feat(ocr): add OCR (#5834)

2026-05-07 12:25:37 +02:00 · 2026-04-03 11:35:36 +03:00
parent 8bd1da0552 3e767b4723
commit 1f0a6b4a79
64 changed files with 2780 additions and 252 deletions
--- a/apps/client/src/components/app_context.ts
+++ b/apps/client/src/components/app_context.ts
@@ -302,6 +302,7 @@ export type CommandMappings = {
    ninthTab: CommandData;
    lastTab: CommandData;
    showNoteSource: CommandData;
+    showNoteOCRText: CommandData;
    showSQLConsole: CommandData;
    showBackendLog: CommandData;
    showCheatsheet: CommandData;
--- a/apps/client/src/components/root_command_executor.ts
+++ b/apps/client/src/components/root_command_executor.ts
@@ -148,6 +148,19 @@ export default class RootCommandExecutor extends Component {
        }
    }

+    async showNoteOCRTextCommand() {
+        const notePath = appContext.tabManager.getActiveContextNotePath();
+
+        if (notePath) {
+            await appContext.tabManager.openTabWithNoteWithHoisting(notePath, {
+                activate: true,
+                viewScope: {
+                    viewMode: "ocr"
+                }
+            });
+        }
+    }
+
    async showAttachmentsCommand() {
        const notePath = appContext.tabManager.getActiveContextNotePath();

--- a/apps/client/src/services/content_renderer.ts
+++ b/apps/client/src/services/content_renderer.ts
@@ -1,6 +1,6 @@
 import "./content_renderer.css";

-import { normalizeMimeTypeForCKEditor } from "@triliumnext/commons";
+import { normalizeMimeTypeForCKEditor, type TextRepresentationResponse } from "@triliumnext/commons";
 import { h, render } from "preact";
 import WheelZoom from 'vanilla-js-wheel-zoom';

@@ -15,6 +15,7 @@ import openService from "./open.js";
 import protectedSessionService from "./protected_session.js";
 import protectedSessionHolder from "./protected_session_holder.js";
 import renderService from "./render.js";
+import server from "./server.js";
 import { applySingleBlockSyntaxHighlight } from "./syntax_highlight.js";
 import utils, { getErrorMessage } from "./utils.js";

@@ -32,6 +33,7 @@ export interface RenderOptions {
    includeArchivedNotes?: boolean;
    /** Set of note IDs that have already been seen during rendering to prevent infinite recursion. */
    seenNoteIds?: Set<string>;
+    showTextRepresentation?: boolean;
 }

 const CODE_MIME_TYPES = new Set(["application/json"]);
@@ -55,9 +57,9 @@ export async function getRenderedContent(this: {} | { ctx: string }, entity: FNo
    } else if (type === "code") {
        await renderCode(entity, $renderedContent);
    } else if (["image", "canvas", "mindMap", "spreadsheet"].includes(type)) {
-        renderImage(entity, $renderedContent, options);
+        await renderImage(entity, $renderedContent, options);
    } else if (!options.tooltip && ["file", "pdf", "audio", "video"].includes(type)) {
-        await renderFile(entity, type, $renderedContent);
+        await renderFile(entity, type, $renderedContent, options);
    } else if (type === "mermaid") {
        await renderMermaid(entity, $renderedContent);
    } else if (type === "render" && entity instanceof FNote) {
@@ -138,7 +140,7 @@ async function renderCode(note: FNote | FAttachment, $renderedContent: JQuery<HT
    await applySingleBlockSyntaxHighlight($codeBlock, normalizeMimeTypeForCKEditor(note.mime));
 }

-function renderImage(entity: FNote | FAttachment, $renderedContent: JQuery<HTMLElement>, options: RenderOptions = {}) {
+async function renderImage(entity: FNote | FAttachment, $renderedContent: JQuery<HTMLElement>, options: RenderOptions = {}) {
    const encodedTitle = encodeURIComponent(entity.title);

    let url;
@@ -146,13 +148,14 @@ function renderImage(entity: FNote | FAttachment, $renderedContent: JQuery<HTMLE
    if (entity instanceof FNote) {
        url = `api/images/${entity.noteId}/${encodedTitle}?${Math.random()}`;
    } else if (entity instanceof FAttachment) {
-        url = `api/attachments/${entity.attachmentId}/image/${encodedTitle}?${entity.utcDateModified}">`;
+        url = `api/attachments/${entity.attachmentId}/image/${encodedTitle}?${entity.utcDateModified}`;
    }

    $renderedContent // styles needed for the zoom to work well
        .css("display", "flex")
        .css("align-items", "center")
-        .css("justify-content", "center");
+        .css("justify-content", "center")
+        .css("flex-direction", "column");   // OCR text is displayed below the image.

    const $img = $("<img>")
        .attr("src", url || "")
@@ -178,9 +181,35 @@ function renderImage(entity: FNote | FAttachment, $renderedContent: JQuery<HTMLE
    }

    imageContextMenuService.setupContextMenu($img);
+
+    if (entity instanceof FNote && options.showTextRepresentation) {
+        await addOCRTextIfAvailable(entity, $renderedContent);
+    }
 }

-async function renderFile(entity: FNote | FAttachment, type: string, $renderedContent: JQuery<HTMLElement>) {
+async function addOCRTextIfAvailable(note: FNote, $content: JQuery<HTMLElement>) {
+    try {
+        const data = await server.get<TextRepresentationResponse>(`ocr/notes/${note.noteId}/text`);
+        if (data.success && data.hasOcr && data.text) {
+            const $ocrSection = $(`
+                <div class="ocr-text-section">
+                    <div class="ocr-header">
+                        <span class="bx bx-text"></span> ${t("ocr.extracted_text")}
+                    </div>
+                    <div class="ocr-content"></div>
+                </div>
+            `);
+
+            $ocrSection.find('.ocr-content').text(data.text);
+            $content.append($ocrSection);
+        }
+    } catch (error) {
+        // Silently fail if OCR API is not available
+        console.debug('Failed to fetch OCR text:', error);
+    }
+}
+
+async function renderFile(entity: FNote | FAttachment, type: string, $renderedContent: JQuery<HTMLElement>, options: RenderOptions = {}) {
    let entityType, entityId;

    if (entity instanceof FNote) {
@@ -220,6 +249,10 @@ async function renderFile(entity: FNote | FAttachment, type: string, $renderedCo
        $content.append($videoPreview);
    }

+    if (entity instanceof FNote && options.showTextRepresentation) {
+        await addOCRTextIfAvailable(entity, $content);
+    }
+
    if (entityType === "notes" && "noteId" in entity) {
        // TODO: we should make this available also for attachments, but there's a problem with "Open externally" support
        //       in attachment list
--- a/apps/client/src/services/link.ts
+++ b/apps/client/src/services/link.ts
@@ -28,7 +28,7 @@ async function getLinkIcon(noteId: string, viewMode: ViewMode | undefined) {
    return icon;
 }

-export type ViewMode = "default" | "source" | "attachments" | "contextual-help" | "note-map";
+export type ViewMode = "default" | "source" | "attachments" | "contextual-help" | "note-map" | "ocr";

 export interface ViewScope {
    /**
--- a/apps/client/src/services/server.ts
+++ b/apps/client/src/services/server.ts
@@ -270,7 +270,11 @@ function ajax(url: string, method: string, data: unknown, headers: Headers, opts
                } else if (opts.silentInternalServerError && jqXhr.status === 500) {
                    // report nothing
                } else {
-                    await reportError(method, url, jqXhr.status, jqXhr.responseText);
+                    try {
+                        await reportError(method, url, jqXhr.status, jqXhr.responseText);
+                    } catch {
+                        // reportError may throw (e.g. ValidationError); ensure rej() is still called below.
+                    }
                }

                rej(jqXhr.responseText);
--- a/apps/client/src/stylesheets/style.css
+++ b/apps/client/src/stylesheets/style.css
@@ -2641,3 +2641,26 @@ iframe.print-iframe {
    min-height: 50px;
    align-items: center;
 }
+
+.ocr-text-section {
+    padding: 10px;
+    background: var(--accented-background-color);
+    border-left: 3px solid var(--main-border-color);
+    text-align: left;
+    width: 100%;
+}
+
+.ocr-header {
+    font-weight: bold;
+    margin-bottom: 8px;
+    font-size: 0.9em;
+    color: var(--muted-text-color);
+}
+
+.ocr-content {
+    max-height: 150px;
+    overflow-y: auto;
+    font-size: 0.9em;
+    line-height: 1.4;
+    white-space: pre-wrap;
+}
--- a/apps/client/src/translations/en/translation.json
+++ b/apps/client/src/translations/en/translation.json
@@ -691,6 +691,7 @@
    "search_in_note": "Search in note",
    "note_source": "Note source",
    "note_attachments": "Note attachments",
+    "view_ocr_text": "View OCR text",
    "open_note_externally": "Open note externally",
    "open_note_externally_title": "File will be open in an external application and watched for changes. You'll then be able to upload the modified version back to Trilium.",
    "open_note_custom": "Open note custom",
@@ -1254,12 +1255,28 @@
  },
  "images": {
    "images_section_title": "Images",
-    "download_images_automatically": "Download images automatically for offline use.",
-    "download_images_description": "Pasted HTML can contain references to online images, Trilium will find those references and download the images so that they are available offline.",
-    "enable_image_compression": "Enable image compression",
-    "max_image_dimensions": "Max width / height of an image (image will be resized if it exceeds this setting).",
+    "download_images_automatically": "Download images automatically",
+    "download_images_description": "Download referenced online images from pasted HTML so they are available offline.",
+    "enable_image_compression": "Image compression",
+    "enable_image_compression_description": "Compress and resize images when they are uploaded or pasted.",
+    "max_image_dimensions": "Max image dimensions",
+    "max_image_dimensions_description": "Images exceeding this size will be resized automatically.",
    "max_image_dimensions_unit": "pixels",
-    "jpeg_quality_description": "JPEG quality (10 - worst quality, 100 - best quality, 50 - 85 is recommended)"
+    "jpeg_quality": "JPEG quality",
+    "jpeg_quality_description": "Recommended range is 50–85. Lower values reduce file size, higher values preserve detail.",
+    "ocr_section_title": "Text Extraction (OCR)",
+    "ocr_related_content_languages": "Content languages (used for text extraction)",
+    "ocr_auto_process": "Auto-process new files",
+    "ocr_auto_process_description": "Automatically extract text from newly uploaded or pasted files.",
+    "ocr_min_confidence": "Minimum confidence",
+    "ocr_confidence_description": "Only extract text above this confidence threshold. Lower values include more text but may be less accurate.",
+    "batch_ocr_title": "Process Existing Files",
+    "batch_ocr_description": "Extract text from all existing images, PDFs, and Office documents in your notes. This may take some time depending on the number of files.",
+    "batch_ocr_start": "Start Batch Processing",
+    "batch_ocr_starting": "Starting batch processing...",
+    "batch_ocr_progress": "Processing {{processed}} of {{total}} files...",
+    "batch_ocr_completed": "Batch processing completed! Processed {{processed}} files.",
+    "batch_ocr_error": "Error during batch processing: {{error}}"
  },
  "attachment_erasure_timeout": {
    "attachment_erasure_timeout": "Attachment Erasure Timeout",
@@ -1967,7 +1984,7 @@
  },
  "content_language": {
    "title": "Content languages",
-    "description": "Select one or more languages that should appear in the language selection in the Basic Properties section of a read-only or editable text note. This will allow features such as spell-checking or right-to-left support."
+    "description": "Select one or more languages that should appear in the language selection in the Basic Properties section of a read-only or editable text note. This will allow features such as spell-checking, right-to-left support and text extraction (OCR)."
  },
  "switch_layout_button": {
    "title_vertical": "Move editing pane to the bottom",
@@ -2067,6 +2084,19 @@
  "calendar_view": {
    "delete_note": "Delete note..."
  },
+  "ocr": {
+    "extracted_text": "Extracted Text (OCR)",
+    "extracted_text_title": "Extracted Text (OCR)",
+    "loading_text": "Loading OCR text...",
+    "no_text_available": "No OCR text available",
+    "no_text_explanation": "This note has not been processed for OCR text extraction or no text was found.",
+    "failed_to_load": "Failed to load OCR text",
+    "process_now": "Process OCR",
+    "processing": "Processing...",
+    "processing_started": "OCR processing has been started. Please wait a moment and refresh.",
+    "processing_failed": "Failed to start OCR processing",
+    "view_extracted_text": "View extracted text (OCR)"
+  },
  "command_palette": {
    "tree-action-name": "Tree: {{name}}",
    "export_note_title": "Export Note",
--- a/apps/client/src/widgets/NoteDetail.tsx
+++ b/apps/client/src/widgets/NoteDetail.tsx
@@ -336,6 +336,8 @@ export async function getExtendedWidgetType(note: FNote | null | undefined, note

    if (noteContext?.viewScope?.viewMode === "source") {
        resultingType = "readOnlyCode";
+    } else if (noteContext.viewScope?.viewMode === "ocr") {
+        resultingType = "readOnlyOCRText";
    } else if (noteContext.viewScope?.viewMode === "attachments") {
        resultingType = noteContext.viewScope.attachmentId ? "attachmentDetail" : "attachmentList";
    } else if (noteContext.viewScope?.viewMode === "note-map") {
--- a/apps/client/src/widgets/collections/NoteList.tsx
+++ b/apps/client/src/widgets/collections/NoteList.tsx
@@ -25,6 +25,7 @@ interface NoteListProps {
    viewType: ViewTypeOptions | undefined;
    onReady?: (data: PrintReport) => void;
    onProgressChanged?(progress: number): void;
+    showTextRepresentation?: boolean;
 }

 type LazyLoadedComponent = ((props: ViewModeProps<any>) => VNode<any> | undefined);
@@ -67,7 +68,7 @@ export default function NoteList(props: Pick<NoteListProps, "displayOnlyCollecti

 export function SearchNoteList(props: Omit<NoteListProps, "isEnabled" | "viewType">) {
    const viewType = useNoteViewType(props.note);
-    return <CustomNoteList {...props} isEnabled={true} viewType={viewType} />;
+    return <CustomNoteList {...props} isEnabled={true} viewType={viewType} showTextRepresentation />;
 }

 export function CustomNoteList({ note, viewType, isEnabled: shouldEnable, notePath, highlightedTokens, displayOnlyCollections, ntxId, onReady, onProgressChanged, ...restProps }: NoteListProps) {
--- a/apps/client/src/widgets/collections/board/data.spec.ts
+++ b/apps/client/src/widgets/collections/board/data.spec.ts
@@ -1,8 +1,9 @@
-import { it, describe, expect } from "vitest";
-import { buildNote } from "../../../test/easy-froca";
-import { getBoardData } from "./data";
+import { describe, expect,it } from "vitest";
+
 import FBranch from "../../../entities/fbranch";
 import froca from "../../../services/froca";
+import { buildNote } from "../../../test/easy-froca";
+import { getBoardData } from "./data";

 describe("Board data", () => {
    it("deduplicates cloned notes", async () => {
--- a/apps/client/src/widgets/collections/interface.ts
+++ b/apps/client/src/widgets/collections/interface.ts
@@ -21,4 +21,5 @@ export interface ViewModeProps<T extends object> {
    media: ViewModeMedia;
    onReady(data: PrintReport): void;
    onProgressChanged?: ProgressChangedFn;
+    showTextRepresentation?: boolean;
 }
--- a/apps/client/src/widgets/collections/legacy/ListOrGridView.tsx
+++ b/apps/client/src/widgets/collections/legacy/ListOrGridView.tsx
@@ -23,7 +23,7 @@ import { ComponentChildren, TargetedMouseEvent } from "preact";

 const contentSizeObserver = new ResizeObserver(onContentResized);

-export function ListView({ note, noteIds: unfilteredNoteIds, highlightedTokens }: ViewModeProps<{}>) {
+export function ListView({ note, noteIds: unfilteredNoteIds, highlightedTokens, showTextRepresentation }: ViewModeProps<{}>) {
    const expandDepth = useExpansionDepth(note);
    const noteIds = useFilteredNoteIds(note, unfilteredNoteIds);
    const { pageNotes, ...pagination } = usePagination(note, noteIds);
@@ -37,13 +37,14 @@ export function ListView({ note, noteIds: unfilteredNoteIds, highlightedTokens }
                    key={childNote.noteId}
                    note={childNote} parentNote={note}
                    expandDepth={expandDepth} highlightedTokens={highlightedTokens}
-                    currentLevel={1} includeArchived={includeArchived} />
+                    currentLevel={1} includeArchived={includeArchived}
+                    showTextRepresentation={showTextRepresentation} />
            ))}
        </Card>
    </NoteList>;
 }

-export function GridView({ note, noteIds: unfilteredNoteIds, highlightedTokens }: ViewModeProps<{}>) {
+export function GridView({ note, noteIds: unfilteredNoteIds, highlightedTokens, showTextRepresentation }: ViewModeProps<{}>) {
    const noteIds = useFilteredNoteIds(note, unfilteredNoteIds);
    const { pageNotes, ...pagination } = usePagination(note, noteIds);
    const [ includeArchived ] = useNoteLabelBoolean(note, "includeArchived");
@@ -56,7 +57,8 @@ export function GridView({ note, noteIds: unfilteredNoteIds, highlightedTokens }
                                note={childNote}
                                parentNote={note}
                                highlightedTokens={highlightedTokens}
-                                includeArchived={includeArchived} />
+                                includeArchived={includeArchived}
+                                showTextRepresentation={showTextRepresentation} />
            ))}
        </div>
    </NoteList>
@@ -91,13 +93,14 @@ function NoteList(props: NoteListProps) {
    </div>
 }

-function ListNoteCard({ note, parentNote, highlightedTokens, currentLevel, expandDepth, includeArchived }: {
+function ListNoteCard({ note, parentNote, highlightedTokens, currentLevel, expandDepth, includeArchived, showTextRepresentation }: {
    note: FNote,
    parentNote: FNote,
    currentLevel: number,
    expandDepth: number,
    highlightedTokens: string[] | null | undefined;
    includeArchived: boolean;
+    showTextRepresentation?: boolean;
 }) {

    const [ isExpanded, setExpanded ] = useState(currentLevel <= expandDepth);
@@ -113,7 +116,8 @@ function ListNoteCard({ note, parentNote, highlightedTokens, currentLevel, expan
                <NoteContent note={note}
                             highlightedTokens={highlightedTokens}
                             noChildrenList
-                             includeArchivedNotes={includeArchived} />
+                             includeArchivedNotes={includeArchived}
+                             showTextRepresentation={showTextRepresentation} />
            </CardSection>

            <NoteChildren note={note}
@@ -157,6 +161,7 @@ interface GridNoteCardProps {
    parentNote: FNote;
    highlightedTokens: string[] | null | undefined;
    includeArchived: boolean;
+    showTextRepresentation?: boolean;
 }

 function GridNoteCard(props: GridNoteCardProps) {
@@ -185,6 +190,7 @@ function GridNoteCard(props: GridNoteCardProps) {
                         trim
                         highlightedTokens={props.highlightedTokens}
                         includeArchivedNotes={props.includeArchived}
+                         showTextRepresentation={props.showTextRepresentation}
            />
        </CardFrame>
    );
@@ -201,12 +207,13 @@ function NoteAttributes({ note }: { note: FNote }) {
    return <span className="note-list-attributes" ref={ref} />;
 }

-export function NoteContent({ note, trim, noChildrenList, highlightedTokens, includeArchivedNotes }: {
+export function NoteContent({ note, trim, noChildrenList, highlightedTokens, includeArchivedNotes, showTextRepresentation }: {
    note: FNote;
    trim?: boolean;
    noChildrenList?: boolean;
    highlightedTokens: string[] | null | undefined;
    includeArchivedNotes: boolean;
+    showTextRepresentation?: boolean;
 }) {
    const contentRef = useRef<HTMLDivElement>(null);
    const highlightSearch = useImperativeSearchHighlighlighting(highlightedTokens);
@@ -230,7 +237,8 @@ export function NoteContent({ note, trim, noChildrenList, highlightedTokens, inc
            trim,
            noChildrenList,
            noIncludedNotes: true,
-            includeArchivedNotes
+            includeArchivedNotes,
+            showTextRepresentation
        })
            .then(({ $renderedContent, type }) => {
                if (!contentRef.current) return;
--- a/apps/client/src/widgets/mobile_widgets/TabSwitcher.tsx
+++ b/apps/client/src/widgets/mobile_widgets/TabSwitcher.tsx
@@ -27,6 +27,7 @@ const VIEW_MODE_ICON_MAPPINGS: Record<Exclude<ViewMode, "default">, string> = {
    "contextual-help": "bx bx-help-circle",
    "note-map": "bx bxs-network-chart",
    attachments: "bx bx-paperclip",
+    ocr: "bx bx-text"
 };

 export default function TabSwitcher() {
--- a/apps/client/src/widgets/note_types.tsx
+++ b/apps/client/src/widgets/note_types.tsx
@@ -12,7 +12,7 @@ import { TypeWidgetProps } from "./type_widgets/type_widget";
 * A `NoteType` altered by the note detail widget, taking into consideration whether the note is editable or not and adding special note types such as an empty one,
 * for protected session or attachment information.
 */
-export type ExtendedNoteType = Exclude<NoteType, "launcher" | "text" | "code" | "llmChat"> | "empty" | "readOnlyCode" | "readOnlyText" | "editableText" | "editableCode" | "attachmentDetail" | "attachmentList" |  "protectedSession" | "sqlConsole" | "llmChat";
+export type ExtendedNoteType = Exclude<NoteType, "launcher" | "text" | "code" | "llmChat"> | "empty" | "readOnlyCode" | "readOnlyText" | "readOnlyOCRText" | "editableText" | "editableCode" | "attachmentDetail" | "attachmentList" |  "protectedSession" | "sqlConsole" | "llmChat";

 export type TypeWidget = ((props: TypeWidgetProps) => VNode | JSX.Element | undefined);
 type NoteTypeView = () => (Promise<{ default: TypeWidget } | TypeWidget> | TypeWidget);
@@ -78,6 +78,11 @@ export const TYPE_MAPPINGS: Record<ExtendedNoteType, NoteTypeMapping> = {
        className: "note-detail-readonly-code",
        printable: true
    },
+    readOnlyOCRText: {
+        view: () => import("./type_widgets/ReadOnlyTextRepresentation"),
+        className: "note-detail-ocr-text",
+        printable: true
+    },
    editableCode: {
        view: async () => (await import("./type_widgets/code/Code")).EditableCode,
        className: "note-detail-code",
--- a/apps/client/src/widgets/react/Slider.tsx
+++ b/apps/client/src/widgets/react/Slider.tsx
@@ -3,6 +3,7 @@ interface SliderProps {
    onChange(newValue: number);
    min?: number;
    max?: number;
+    step?: number;
    title?: string;
 }

--- a/apps/client/src/widgets/ribbon/NoteActions.tsx
+++ b/apps/client/src/widgets/ribbon/NoteActions.tsx
@@ -162,6 +162,7 @@ export function NoteContextMenu({ note, noteContext, itemsAtStart, itemsNearNote
                <CommandItem command="openNoteExternally" icon="bx bx-file-find" disabled={isSearchOrBook || !isElectron} text={t("note_actions.open_note_externally")} title={t("note_actions.open_note_externally_title")} />
                <CommandItem command="openNoteCustom" icon="bx bx-customize" disabled={isSearchOrBook || isMac || !isElectron} text={t("note_actions.open_note_custom")} />
                <CommandItem command="showNoteSource" icon="bx bx-code" disabled={!hasSource} text={t("note_actions.note_source")} />
+                <CommandItem command="showNoteOCRText" icon="bx bx-text" disabled={!["image", "file"].includes(noteType)} text={t("note_actions.view_ocr_text")} />
                {(syncServerHost && isElectron) &&
                    <CommandItem command="openNoteOnServer" icon="bx bx-world" disabled={!syncServerHost} text={t("note_actions.open_note_on_server")} />
                }
--- a/apps/client/src/widgets/type_widgets/Attachment.tsx
+++ b/apps/client/src/widgets/type_widgets/Attachment.tsx
@@ -27,8 +27,10 @@ import { FormDropdownDivider, FormListItem } from "../react/FormList";
 import HelpButton from "../react/HelpButton";
 import { useTriliumEvent } from "../react/hooks";
 import Icon from "../react/Icon";
+import Modal from "../react/Modal";
 import NoteLink from "../react/NoteLink";
 import { ParentComponent, refToJQuerySelector } from "../react/react_utils";
+import { TextRepresentation } from "./ReadOnlyTextRepresentation";
 import { TypeWidgetProps } from "./type_widget";

 /**
@@ -141,6 +143,8 @@ export function AttachmentDetail({ note, viewScope }: TypeWidgetProps) {

 function AttachmentInfo({ attachment, isFullDetail }: { attachment: FAttachment, isFullDetail?: boolean }) {
    const contentWrapper = useRef<HTMLDivElement>(null);
+    const [ ocrModalShown, setOcrModalShown ] = useState(false);
+    const supportsOcr = attachment.role === "image" || attachment.role === "file";

    function refresh() {
        content_renderer.getRenderedContent(attachment, { imageHasZoom: isFullDetail })
@@ -181,7 +185,11 @@ function AttachmentInfo({ attachment, isFullDetail }: { attachment: FAttachment,
        <div className="attachment-detail-widget">
            <div className={`attachment-detail-wrapper ${isFullDetail ? "full-detail" : "list-view"} ${attachment.utcDateScheduledForErasureSince ? "scheduled-for-deletion" : ""}`}>
                <div className="attachment-title-line">
-                    <AttachmentActions attachment={attachment} copyAttachmentLinkToClipboard={copyAttachmentLinkToClipboard} />
+                    <AttachmentActions
+                        attachment={attachment}
+                        copyAttachmentLinkToClipboard={copyAttachmentLinkToClipboard}
+                        onShowOcr={supportsOcr ? () => setOcrModalShown(true) : undefined}
+                    />
                    <h4 className="attachment-title">
                        {!isFullDetail ? (
                            <NoteLink
@@ -207,6 +215,22 @@ function AttachmentInfo({ attachment, isFullDetail }: { attachment: FAttachment,
                {attachment.utcDateScheduledForErasureSince && <DeletionAlert utcDateScheduledForErasureSince={attachment.utcDateScheduledForErasureSince} />}
                <div ref={contentWrapper} className="attachment-content-wrapper" />
            </div>
+
+            {supportsOcr && (
+                <Modal
+                    className="ocr-text-modal"
+                    title={t("ocr.extracted_text_title")}
+                    show={ocrModalShown}
+                    onHidden={() => setOcrModalShown(false)}
+                    size="lg"
+                    scrollable
+                >
+                    <TextRepresentation
+                        textUrl={`ocr/attachments/${attachment.attachmentId}/text`}
+                        processUrl={`ocr/process-attachment/${attachment.attachmentId}`}
+                    />
+                </Modal>
+            )}
        </div>
    );
 }
@@ -228,7 +252,7 @@ function DeletionAlert({ utcDateScheduledForErasureSince }: { utcDateScheduledFo
    );
 }

-function AttachmentActions({ attachment, copyAttachmentLinkToClipboard }: { attachment: FAttachment, copyAttachmentLinkToClipboard: () => void }) {
+function AttachmentActions({ attachment, copyAttachmentLinkToClipboard, onShowOcr }: { attachment: FAttachment, copyAttachmentLinkToClipboard: () => void, onShowOcr?: () => void }) {
    const isElectron = utils.isElectron();
    const fileUploadRef = useRef<HTMLInputElement>(null);

@@ -262,6 +286,12 @@ function AttachmentActions({ attachment, copyAttachmentLinkToClipboard }: { atta
                    icon="bx bx-link"
                    onClick={copyAttachmentLinkToClipboard}
                >{t("attachments_actions.copy_link_to_clipboard")}</FormListItem>
+                {onShowOcr && (
+                    <FormListItem
+                        icon="bx bx-text"
+                        onClick={onShowOcr}
+                    >{t("ocr.view_extracted_text")}</FormListItem>
+                )}
                <FormDropdownDivider />

                <FormListItem
--- a/apps/client/src/widgets/type_widgets/ContentWidget.tsx
+++ b/apps/client/src/widgets/type_widgets/ContentWidget.tsx
@@ -4,7 +4,7 @@ import AppearanceSettings from "./options/appearance";
 import ShortcutSettings from "./options/shortcuts";
 import TextNoteSettings from "./options/text_notes";
 import CodeNoteSettings from "./options/code_notes";
-import ImageSettings from "./options/images";
+import MediaSettings from "./options/media";
 import SpellcheckSettings from "./options/spellcheck";
 import PasswordSettings from "./options/password";
 import MultiFactorAuthenticationSettings from "./options/multi_factor_authentication";
@@ -19,14 +19,14 @@ import "./ContentWidget.css";
 import { t } from "../../services/i18n";
 import BackendLog from "./code/BackendLog";

-export type OptionPages = "_optionsAppearance" | "_optionsShortcuts" | "_optionsTextNotes" | "_optionsCodeNotes" | "_optionsImages" | "_optionsSpellcheck" | "_optionsPassword" | "_optionsMFA" | "_optionsEtapi" | "_optionsBackup" | "_optionsSync" | "_optionsOther" | "_optionsLocalization" | "_optionsAdvanced" | "_optionsLlm";
+export type OptionPages = "_optionsAppearance" | "_optionsShortcuts" | "_optionsTextNotes" | "_optionsCodeNotes" | "_optionsMedia" | "_optionsSpellcheck" | "_optionsPassword" | "_optionsMFA" | "_optionsEtapi" | "_optionsBackup" | "_optionsSync" | "_optionsOther" | "_optionsLocalization" | "_optionsAdvanced" | "_optionsLlm";

 const CONTENT_WIDGETS: Record<OptionPages | "_backendLog", (props: TypeWidgetProps) => JSX.Element> = {
    _optionsAppearance: AppearanceSettings,
    _optionsShortcuts: ShortcutSettings,
    _optionsTextNotes: TextNoteSettings,
    _optionsCodeNotes: CodeNoteSettings,
-    _optionsImages: ImageSettings,
+    _optionsMedia: MediaSettings,
    _optionsSpellcheck: SpellcheckSettings,
    _optionsPassword: PasswordSettings,
    _optionsMFA: MultiFactorAuthenticationSettings,
--- a/apps/client/src/widgets/type_widgets/ReadOnlyTextRepresentation.css
+++ b/apps/client/src/widgets/type_widgets/ReadOnlyTextRepresentation.css
@@ -0,0 +1,56 @@
+.text-representation {
+    padding: 10px;
+}
+
+.text-representation-header {
+    margin-bottom: 10px;
+    padding: 8px 12px;
+    background-color: var(--main-background-color);
+    border: 1px solid var(--main-border-color);
+    border-radius: 4px;
+    font-weight: 500;
+}
+
+.text-representation-loading {
+    text-align: center;
+    padding: 30px;
+    color: var(--muted-text-color);
+}
+
+.text-representation-content {
+    white-space: pre-wrap;
+    line-height: 1.6;
+    border: 1px solid var(--main-border-color);
+    border-radius: 4px;
+    padding: 15px;
+    background-color: var(--accented-background-color);
+    min-height: 100px;
+    user-select: text;
+}
+
+.text-representation-meta {
+    font-size: 0.9em;
+    color: var(--muted-text-color);
+    margin-top: 10px;
+    font-style: italic;
+}
+
+.text-representation-empty {
+    color: var(--muted-text-color);
+    font-style: italic;
+    text-align: center;
+    padding: 30px;
+}
+
+.text-representation-process-btn {
+    margin-top: 15px;
+}
+
+.text-representation-error {
+    color: var(--error-color);
+    background-color: var(--error-background-color);
+    border: 1px solid var(--error-border-color);
+    padding: 10px;
+    border-radius: 4px;
+    margin-top: 10px;
+}
--- a/apps/client/src/widgets/type_widgets/ReadOnlyTextRepresentation.tsx
+++ b/apps/client/src/widgets/type_widgets/ReadOnlyTextRepresentation.tsx
@@ -0,0 +1,131 @@
+import "./ReadOnlyTextRepresentation.css";
+
+import type { TextRepresentationResponse } from "@triliumnext/commons";
+import { useEffect, useState } from "preact/hooks";
+
+import { t } from "../../services/i18n";
+import server from "../../services/server";
+import toast from "../../services/toast";
+import { TypeWidgetProps } from "./type_widget";
+
+type State =
+    | { kind: "loading" }
+    | { kind: "loaded"; text: string }
+    | { kind: "empty" }
+    | { kind: "error"; message: string };
+
+interface TextRepresentationProps {
+    /** The API path to fetch OCR text from (e.g. `ocr/notes/{id}/text`). */
+    textUrl: string;
+    /** The API path to trigger OCR processing (e.g. `ocr/process-note/{id}`). */
+    processUrl: string;
+}
+
+export default function ReadOnlyTextRepresentation({ note }: TypeWidgetProps) {
+    return (
+        <TextRepresentation
+            textUrl={`ocr/notes/${note.noteId}/text`}
+            processUrl={`ocr/process-note/${note.noteId}`}
+        />
+    );
+}
+
+export function TextRepresentation({ textUrl, processUrl }: TextRepresentationProps) {
+    const [ state, setState ] = useState<State>({ kind: "loading" });
+    const [ processing, setProcessing ] = useState(false);
+
+    async function fetchText() {
+        setState({ kind: "loading" });
+
+        try {
+            const response = await server.get<TextRepresentationResponse>(textUrl);
+
+            if (!response.success) {
+                setState({ kind: "error", message: response.message || t("ocr.failed_to_load") });
+                return;
+            }
+
+            if (!response.hasOcr || !response.text) {
+                setState({ kind: "empty" });
+                return;
+            }
+
+            setState({ kind: "loaded", text: response.text });
+        } catch (error: any) {
+            console.error("Error loading text representation:", error);
+            setState({ kind: "error", message: error.message || t("ocr.failed_to_load") });
+        }
+    }
+
+    useEffect(() => { fetchText(); }, [ textUrl ]);
+
+    async function processOCR() {
+        setProcessing(true);
+        try {
+            const response = await server.post<{ success: boolean; message?: string }>(processUrl, { forceReprocess: true });
+            if (response.success) {
+                toast.showMessage(t("ocr.processing_started"));
+                setTimeout(fetchText, 2000);
+            } else {
+                toast.showError(response.message || t("ocr.processing_failed"));
+            }
+        } catch {
+            // Server errors (4xx/5xx) are already shown as toasts by server.ts.
+        } finally {
+            setProcessing(false);
+        }
+    }
+
+    return (
+        <div className="text-representation note-detail-printable">
+            <div className="text-representation-header">
+                <span className="bx bx-text" />{" "}{t("ocr.extracted_text_title")}
+            </div>
+
+            {state.kind === "loading" && (
+                <div className="text-representation-loading">
+                    <span className="bx bx-loader-alt bx-spin" />{" "}{t("ocr.loading_text")}
+                </div>
+            )}
+
+            {state.kind === "loaded" && (
+                <>
+                    <div className="text-representation-content">
+                        {state.text}
+                    </div>
+                </>
+            )}
+
+            {state.kind === "empty" && (
+                <>
+                    <div className="text-representation-empty">
+                        <span className="bx bx-info-circle" />{" "}{t("ocr.no_text_available")}
+                    </div>
+                    <div className="text-representation-meta">
+                        {t("ocr.no_text_explanation")}
+                    </div>
+                </>
+            )}
+
+            {state.kind === "error" && (
+                <div className="text-representation-error">
+                    <span className="bx bx-error" />{" "}{state.message}
+                </div>
+            )}
+
+            {state.kind !== "loading" && (
+                <button
+                    type="button"
+                    className="btn btn-secondary text-representation-process-btn"
+                    disabled={processing}
+                    onClick={processOCR}
+                >
+                    {processing
+                        ? <><span className="bx bx-loader-alt bx-spin" />{" "}{t("ocr.processing")}</>
+                        : <><span className="bx bx-play" />{" "}{t("ocr.process_now")}</>
+                    }
+                </button>
+            )}
+        </div>
+    );
+}
--- a/apps/client/src/widgets/type_widgets/options/components/OptionsRow.css
+++ b/apps/client/src/widgets/type_widgets/options/components/OptionsRow.css
@@ -1,30 +1,41 @@
 .option-row {
    border-bottom: 1px solid var(--main-border-color);
    display: flex;
-    align-items: center;
+    flex-direction: column;
    padding: 0.5em 0;
 }

-.option-row > label {
-    width: 40%;
+.option-row-main {
+    display: flex;
+    align-items: center;
+}
+
+.option-row-main > label {
+    width: 45%;
    margin-bottom: 0 !important;
    flex-shrink: 0;
 }

-.option-row > select,
-.option-row > .dropdown {
+.option-row-main > select,
+.option-row-main > .dropdown {
    width: 60%;
 }

-.option-row > .dropdown button {
+.option-row-main > .dropdown button {
    width: 100%;
    text-align: start;
 }

+.option-row-description {
+    line-height: 1.3;
+    margin-top: 0.25em;
+    color: var(--muted-text-color);
+}
+
 .option-row:last-of-type {
    border-bottom: unset;
 }

-.option-row.centered {
+.option-row.centered .option-row-main {
    justify-content: center;
 }
--- a/apps/client/src/widgets/type_widgets/options/components/OptionsRow.tsx
+++ b/apps/client/src/widgets/type_widgets/options/components/OptionsRow.tsx
@@ -5,18 +5,22 @@ import { useUniqueName } from "../../../react/hooks";
 interface OptionsRowProps {
    name: string;
    label?: string;
+    description?: string;
    children: VNode;
    centered?: boolean;
 }

-export default function OptionsRow({ name, label, children, centered }: OptionsRowProps) {
+export default function OptionsRow({ name, label, description, children, centered }: OptionsRowProps) {
    const id = useUniqueName(name);
    const childWithId = cloneElement(children, { id });

    return (
        <div className={`option-row ${centered ? "centered" : ""}`}>
-            {label && <label for={id}>{label}</label>}
-            {childWithId}
+            <div className="option-row-main">
+                {label && <label for={id}>{label}</label>}
+                {childWithId}
+            </div>
+            {description && <small className="option-row-description">{description}</small>}
        </div>
    );
 }
--- a/apps/client/src/widgets/type_widgets/options/images.tsx
+++ b/apps/client/src/widgets/type_widgets/options/images.tsx
@@ -1,48 +0,0 @@
-import { t } from "../../../services/i18n";
-import FormCheckbox from "../../react/FormCheckbox";
-import FormGroup from "../../react/FormGroup";
-import { FormTextBoxWithUnit } from "../../react/FormTextBox";
-import { useTriliumOption, useTriliumOptionBool } from "../../react/hooks";
-import OptionsSection from "./components/OptionsSection";
-
-export default function ImageSettings() {
-    const [ downloadImagesAutomatically, setDownloadImagesAutomatically ] = useTriliumOptionBool("downloadImagesAutomatically");
-    const [ compressImages, setCompressImages ] = useTriliumOptionBool("compressImages");
-    const [ imageMaxWidthHeight, setImageMaxWidthHeight ] = useTriliumOption("imageMaxWidthHeight"); 
-    const [ imageJpegQuality, setImageJpegQuality ] = useTriliumOption("imageJpegQuality");
-
-    return (
-        <OptionsSection title={t("images.images_section_title")}>
-            <FormGroup name="download-images-automatically" description={t("images.download_images_description")}>
-                <FormCheckbox                    
-                    label={t("images.download_images_automatically")}
-                    currentValue={downloadImagesAutomatically} onChange={setDownloadImagesAutomatically}
-                />
-            </FormGroup>
-
-            <hr/>
-
-            <FormCheckbox
-                name="image-compression-enabled"
-                label={t("images.enable_image_compression")}
-                currentValue={compressImages} onChange={setCompressImages}
-            />
-
-            <FormGroup name="image-max-width-height" label={t("images.max_image_dimensions")} disabled={!compressImages}>
-                <FormTextBoxWithUnit
-                    type="number" min="1"
-                    unit={t("images.max_image_dimensions_unit")}
-                    currentValue={imageMaxWidthHeight} onChange={setImageMaxWidthHeight}
-                />
-            </FormGroup>
-
-            <FormGroup name="image-jpeg-quality" label={t("images.jpeg_quality_description")} disabled={!compressImages}>
-                <FormTextBoxWithUnit                    
-                    min="10" max="100" type="number"
-                    unit={t("units.percentage")}
-                    currentValue={imageJpegQuality} onChange={setImageJpegQuality}
-                />
-            </FormGroup>
-        </OptionsSection>
-    );
-}
--- a/apps/client/src/widgets/type_widgets/options/media.tsx
+++ b/apps/client/src/widgets/type_widgets/options/media.tsx
@@ -0,0 +1,176 @@
+import { useCallback, useEffect, useRef, useState } from "preact/hooks";
+
+import { t } from "../../../services/i18n";
+import server from "../../../services/server";
+import toast from "../../../services/toast";
+import { FormTextBoxWithUnit } from "../../react/FormTextBox";
+import FormToggle from "../../react/FormToggle";
+import { useTriliumOption, useTriliumOptionBool } from "../../react/hooks";
+import Slider from "../../react/Slider";
+import OptionsRow from "./components/OptionsRow";
+import OptionsSection from "./components/OptionsSection";
+import RelatedSettings from "./components/RelatedSettings";
+
+export default function MediaSettings() {
+    return (
+        <>
+            <ImageSettings />
+            <OcrSettings />
+        </>
+    );
+}
+
+function ImageSettings() {
+    const [ downloadImagesAutomatically, setDownloadImagesAutomatically ] = useTriliumOptionBool("downloadImagesAutomatically");
+    const [ compressImages, setCompressImages ] = useTriliumOptionBool("compressImages");
+    const [ imageMaxWidthHeight, setImageMaxWidthHeight ] = useTriliumOption("imageMaxWidthHeight");
+    const [ imageJpegQuality, setImageJpegQuality ] = useTriliumOption("imageJpegQuality");
+
+    return (
+        <OptionsSection title={t("images.images_section_title")}>
+            <OptionsRow name="download-images-automatically" label={t("images.download_images_automatically")} description={t("images.download_images_description")}>
+                <FormToggle
+                    switchOnName="" switchOffName=""
+                    currentValue={downloadImagesAutomatically}
+                    onChange={setDownloadImagesAutomatically}
+                />
+            </OptionsRow>
+
+            <OptionsRow name="image-compression-enabled" label={t("images.enable_image_compression")} description={t("images.enable_image_compression_description")}>
+                <FormToggle
+                    switchOnName="" switchOffName=""
+                    currentValue={compressImages}
+                    onChange={setCompressImages}
+                />
+            </OptionsRow>
+
+            <OptionsRow name="image-max-width-height" label={t("images.max_image_dimensions")} description={t("images.max_image_dimensions_description")}>
+                <FormTextBoxWithUnit
+                    type="number" min="1"
+                    disabled={!compressImages}
+                    unit={t("images.max_image_dimensions_unit")}
+                    currentValue={imageMaxWidthHeight} onChange={setImageMaxWidthHeight}
+                />
+            </OptionsRow>
+
+            <OptionsRow name="image-jpeg-quality" label={`${t("images.jpeg_quality")} (${imageJpegQuality ?? 75}%)`} description={t("images.jpeg_quality_description")}>
+                <Slider
+                    min={10} max={100} step={5}
+                    value={parseInt(imageJpegQuality ?? "75", 10)}
+                    onChange={(v) => setImageJpegQuality(String(v))}
+                />
+            </OptionsRow>
+        </OptionsSection>
+    );
+}
+
+function OcrSettings() {
+    const [ ocrAutoProcess, setOcrAutoProcess ] = useTriliumOptionBool("ocrAutoProcessImages");
+    const [ ocrMinConfidence, setOcrMinConfidence ] = useTriliumOption("ocrMinConfidence");
+
+    return (
+        <>
+            <OptionsSection title={t("images.ocr_section_title")}>
+                <OptionsRow name="ocr-auto-process" label={t("images.ocr_auto_process")} description={t("images.ocr_auto_process_description")}>
+                    <FormToggle
+                        switchOnName="" switchOffName=""
+                        currentValue={ocrAutoProcess}
+                        onChange={setOcrAutoProcess}
+                    />
+                </OptionsRow>
+
+                <OptionsRow name="ocr-min-confidence" label={`${t("images.ocr_min_confidence")} (${Math.round(parseFloat(ocrMinConfidence ?? "0.75") * 100)}%)`} description={t("images.ocr_confidence_description")}>
+                    <Slider
+                        min={0} max={100} step={5}
+                        value={Math.round(parseFloat(ocrMinConfidence ?? "0.75") * 100)}
+                        onChange={(v) => setOcrMinConfidence(String(v / 100))}
+                    />
+                </OptionsRow>
+
+                <BatchProcessing />
+            </OptionsSection>
+
+            <RelatedSettings items={[
+                {
+                    title: t("images.ocr_related_content_languages"),
+                    targetPage: "_optionsLocalization"
+                }
+            ]} />
+        </>
+    );
+}
+
+interface BatchProgress {
+    inProgress: boolean;
+    total: number;
+    processed: number;
+    percentage?: number;
+}
+
+function BatchProcessing() {
+    const [ progress, setProgress ] = useState<BatchProgress | null>(null);
+    const pollingRef = useRef<ReturnType<typeof setInterval>>(null);
+
+    const pollProgress = useCallback(() => {
+        server.get<BatchProgress>("ocr/batch-progress").then((data) => {
+            setProgress(data);
+            if (!data.inProgress && pollingRef.current) {
+                clearInterval(pollingRef.current);
+                pollingRef.current = null;
+                toast.showMessage(t("images.batch_ocr_completed", { processed: data.processed }));
+            }
+        });
+    }, []);
+
+    // Clean up polling on unmount.
+    useEffect(() => {
+        return () => {
+            if (pollingRef.current) {
+                clearInterval(pollingRef.current);
+            }
+        };
+    }, []);
+
+    async function startBatch() {
+        try {
+            const result = await server.post<{ success: boolean; message?: string }>("ocr/batch-process");
+            if (result.success) {
+                toast.showMessage(t("images.batch_ocr_starting"));
+                pollingRef.current = setInterval(pollProgress, 2000);
+                pollProgress();
+            } else {
+                toast.showError(result.message || t("images.batch_ocr_error", { error: "Unknown" }));
+            }
+        } catch {
+            // Server errors are already shown as toasts by server.ts.
+        }
+    }
+
+    const isRunning = progress?.inProgress ?? false;
+
+    return (
+        <OptionsRow name="batch-ocr" label={t("images.batch_ocr_title")} description={t("images.batch_ocr_description")}>
+            {isRunning ? (
+                <div style={{ width: "100%" }}>
+                    <div className="progress" style={{ height: "24px" }}>
+                        <div
+                            className="progress-bar progress-bar-striped progress-bar-animated"
+                            role="progressbar"
+                            style={{ width: `${progress?.percentage ?? 0}%` }}
+                        >
+                            {t("images.batch_ocr_progress", { processed: progress?.processed ?? 0, total: progress?.total ?? 0 })}
+                        </div>
+                    </div>
+                </div>
+            ) : (
+                <button
+                    type="button"
+                    className="btn btn-secondary"
+                    onClick={startBatch}
+                >
+                    <span className="bx bx-play" />{" "}{t("images.batch_ocr_start")}
+                </button>
+            )}
+        </OptionsRow>
+    );
+}
--- a/apps/server/package.json
+++ b/apps/server/package.json
@@ -37,7 +37,8 @@
    "better-sqlite3": "12.8.0",
    "html-to-text": "9.0.5",
    "node-html-parser": "7.1.0",
-    "sucrase": "3.35.1"
+    "sucrase": "3.35.1",
+    "unpdf": "1.4.0"
  },
  "devDependencies": {
    "@braintree/sanitize-url": "7.1.2",
@@ -115,6 +116,7 @@
    "mime-types": "3.0.2",
    "multer": "2.1.1",
    "normalize-strings": "1.1.1",
+    "officeparser": "6.0.7",
    "rand-token": "1.0.1",
    "safe-compare": "1.1.4",
    "sanitize-filename": "1.6.4",
@@ -126,6 +128,7 @@
    "striptags": "3.2.0",
    "supertest": "7.2.2",
    "swagger-jsdoc": "6.2.8",
+    "tesseract.js": "6.0.1",
    "time2fa": "1.4.2",
    "tmp": "0.2.5",
    "turnish": "1.8.0",
--- a/apps/server/spec/db/document.db
+++ b/apps/server/spec/db/document.db
--- a/apps/server/src/assets/db/schema.sql
+++ b/apps/server/src/assets/db/schema.sql
@@ -107,6 +107,7 @@ CREATE TABLE IF NOT EXISTS "recent_notes"
 CREATE TABLE IF NOT EXISTS "blobs" (
                                               `blobId`	TEXT NOT NULL,
                                               `content`	TEXT NULL DEFAULT NULL,
+                                               `textRepresentation` TEXT DEFAULT NULL,
                                               `dateModified` TEXT NOT NULL,
                                               `utcDateModified` TEXT NOT NULL,
                                               PRIMARY KEY(`blobId`)
--- a/apps/server/src/assets/translations/en/server.json
+++ b/apps/server/src/assets/translations/en/server.json
@@ -344,7 +344,7 @@
    "shortcuts-title": "Shortcuts",
    "text-notes": "Text Notes",
    "code-notes-title": "Code Notes",
-    "images-title": "Images",
+    "images-title": "Media",
    "spellcheck-title": "Spellcheck",
    "password-title": "Password",
    "multi-factor-authentication-title": "MFA",
--- a/apps/server/src/becca/entities/bblob.ts
+++ b/apps/server/src/becca/entities/bblob.ts
@@ -15,6 +15,7 @@ class BBlob extends AbstractBeccaEntity<BBlob> {

    content!: string | Buffer;
    contentLength!: number;
+    textRepresentation?: string | null;

    constructor(row: BlobRow) {
        super();
@@ -25,6 +26,7 @@ class BBlob extends AbstractBeccaEntity<BBlob> {
        this.blobId = row.blobId;
        this.content = row.content;
        this.contentLength = row.contentLength;
+        this.textRepresentation = row.textRepresentation;
        this.dateModified = row.dateModified;
        this.utcDateModified = row.utcDateModified;
    }
@@ -34,10 +36,16 @@ class BBlob extends AbstractBeccaEntity<BBlob> {
            blobId: this.blobId,
            content: this.content || null,
            contentLength: this.contentLength,
+            textRepresentation: this.textRepresentation || null,
            dateModified: this.dateModified,
            utcDateModified: this.utcDateModified
        };
    }
+
+    protected getPojoToSave() {
+        const { contentLength: _, ...pojo } = this.getPojo();
+        return pojo;
+    }
 }

 export default BBlob;
--- a/apps/server/src/migrations/migrations.ts
+++ b/apps/server/src/migrations/migrations.ts
@@ -6,6 +6,13 @@

 // Migrations should be kept in descending order, so the latest migration is first.
 const MIGRATIONS: (SqlMigration | JsMigration)[] = [
+    // Add text representation column to blobs table
+    {
+        version: 236,
+        sql: /*sql*/`\
+            ALTER TABLE blobs ADD COLUMN textRepresentation TEXT DEFAULT NULL;
+        `
+    },
    // Add missing database indices for query performance
    {
        version: 235,
--- a/apps/server/src/routes/api/ocr.spec.ts
+++ b/apps/server/src/routes/api/ocr.spec.ts
@@ -0,0 +1,56 @@
+import { describe, expect, it, vi, beforeEach } from "vitest";
+import ocrRoutes from "./ocr.js";
+
+// Mock the OCR service
+vi.mock("../../services/ocr/ocr_service.js", () => ({
+    default: {
+        startBatchProcessing: vi.fn(() => Promise.resolve({ success: true })),
+        getBatchProgress: vi.fn(() => ({ inProgress: false, total: 0, processed: 0 }))
+    }
+}));
+
+// Mock becca
+vi.mock("../../becca/becca.js", () => ({
+    default: {}
+}));
+
+// Mock sql
+vi.mock("../../services/sql.js", () => ({
+    default: {
+        getRow: vi.fn()
+    }
+}));
+
+// Mock log
+vi.mock("../../services/log.js", () => ({
+    default: {
+        error: vi.fn()
+    }
+}));
+
+describe("OCR API", () => {
+    beforeEach(() => {
+        vi.clearAllMocks();
+    });
+
+    it("should return success for batch processing", async () => {
+        const result = await ocrRoutes.batchProcessOCR();
+        expect(result).toEqual({ success: true });
+    });
+
+    it("should return batch progress", async () => {
+        const result = await ocrRoutes.getBatchProgress();
+        expect(result).toEqual({ inProgress: false, total: 0, processed: 0 });
+    });
+
+    it("should return 400 when batch processing fails", async () => {
+        const ocrService = await import("../../services/ocr/ocr_service.js");
+        vi.mocked(ocrService.default.startBatchProcessing).mockResolvedValueOnce({
+            success: false,
+            message: "No images found that need OCR processing"
+        });
+
+        const result = await ocrRoutes.batchProcessOCR();
+        expect(result).toEqual([400, { success: false, message: "No images found that need OCR processing" }]);
+    });
+});
--- a/apps/server/src/routes/api/ocr.ts
+++ b/apps/server/src/routes/api/ocr.ts
@@ -0,0 +1,241 @@
+import { TextRepresentationResponse } from "@triliumnext/commons";
+import type { Request } from "express";
+
+import becca from "../../becca/becca.js";
+import ocrService from "../../services/ocr/ocr_service.js";
+import sql from "../../services/sql.js";
+
+/**
+ * @swagger
+ * /api/ocr/process-note/{noteId}:
+ *   post:
+ *     summary: Process OCR for a specific note
+ *     operationId: ocr-process-note
+ *     parameters:
+ *       - name: noteId
+ *         in: path
+ *         required: true
+ *         schema:
+ *           type: string
+ *         description: ID of the note to process
+ *     requestBody:
+ *       required: false
+ *       content:
+ *         application/json:
+ *           schema:
+ *             type: object
+ *             properties:
+ *               language:
+ *                 type: string
+ *                 description: >
+ *                   Tesseract language code to use (e.g. 'eng', 'fra', 'deu', 'eng+fra').
+ *                   If omitted, the language is resolved automatically from the note's language label,
+ *                   the enabled content languages, or the UI locale.
+ *               forceReprocess:
+ *                 type: boolean
+ *                 description: Force reprocessing even if OCR already exists
+ *                 default: false
+ *     responses:
+ *       '200':
+ *         description: OCR processing completed successfully
+ *       '400':
+ *         description: Bad request - unsupported file type
+ *       '404':
+ *         description: Note not found
+ *       '500':
+ *         description: Internal server error
+ *     security:
+ *       - session: []
+ *     tags: ["ocr"]
+ */
+async function processNoteOCR(req: Request<{ noteId: string }>) {
+    const { noteId } = req.params;
+    const { language, forceReprocess = false } = req.body || {};
+
+    const note = becca.getNote(noteId);
+    if (!note) {
+        return [404, { success: false, message: 'Note not found' }];
+    }
+
+    const result = await ocrService.processNoteOCR(noteId, { language, forceReprocess });
+    if (!result) {
+        return [400, { success: false, message: 'Note is not an image or has unsupported format' }];
+    }
+
+    return { success: true, result };
+}
+
+/**
+ * @swagger
+ * /api/ocr/process-attachment/{attachmentId}:
+ *   post:
+ *     summary: Process OCR for a specific attachment
+ *     operationId: ocr-process-attachment
+ *     parameters:
+ *       - name: attachmentId
+ *         in: path
+ *         required: true
+ *         schema:
+ *           type: string
+ *         description: ID of the attachment to process
+ *     requestBody:
+ *       required: false
+ *       content:
+ *         application/json:
+ *           schema:
+ *             type: object
+ *             properties:
+ *               language:
+ *                 type: string
+ *                 description: >
+ *                   Tesseract language code to use (e.g. 'eng', 'fra', 'deu', 'eng+fra').
+ *                   If omitted, the language is resolved automatically from the owner note's language label,
+ *                   the enabled content languages, or the UI locale.
+ *               forceReprocess:
+ *                 type: boolean
+ *                 description: Force reprocessing even if OCR already exists
+ *                 default: false
+ *     responses:
+ *       '200':
+ *         description: OCR processing completed successfully
+ *       '400':
+ *         description: Bad request - unsupported file type
+ *       '404':
+ *         description: Attachment not found
+ *       '500':
+ *         description: Internal server error
+ *     security:
+ *       - session: []
+ *     tags: ["ocr"]
+ */
+async function processAttachmentOCR(req: Request<{ attachmentId: string }>) {
+    const { attachmentId } = req.params;
+    const { language, forceReprocess = false } = req.body || {};
+
+    const attachment = becca.getAttachment(attachmentId);
+    if (!attachment) {
+        return [404, { success: false, message: 'Attachment not found' }];
+    }
+
+    const result = await ocrService.processAttachmentOCR(attachmentId, { language, forceReprocess });
+    if (!result) {
+        return [400, { success: false, message: 'Attachment is not an image or has unsupported format' }];
+    }
+
+    return { success: true, result };
+}
+
+/**
+ * @swagger
+ * /api/ocr/batch-process:
+ *   post:
+ *     summary: Process OCR for all images without existing OCR results
+ *     operationId: ocr-batch-process
+ *     responses:
+ *       '200':
+ *         description: Batch processing initiated successfully
+ *       '400':
+ *         description: Bad request - OCR disabled or already processing
+ *       '500':
+ *         description: Internal server error
+ *     security:
+ *       - session: []
+ *     tags: ["ocr"]
+ */
+async function batchProcessOCR() {
+    const result = await ocrService.startBatchProcessing();
+    if (!result.success) {
+        return [400, result];
+    }
+    return result;
+}
+
+/**
+ * @swagger
+ * /api/ocr/batch-progress:
+ *   get:
+ *     summary: Get batch OCR processing progress
+ *     operationId: ocr-batch-progress
+ *     responses:
+ *       '200':
+ *         description: Batch processing progress information
+ *       '500':
+ *         description: Internal server error
+ *     security:
+ *       - session: []
+ *     tags: ["ocr"]
+ */
+async function getBatchProgress() {
+    return ocrService.getBatchProgress();
+}
+
+/**
+ * @swagger
+ * /api/ocr/notes/{noteId}/text:
+ *   get:
+ *     summary: Get OCR text for a specific note
+ *     operationId: ocr-get-note-text
+ *     parameters:
+ *       - name: noteId
+ *         in: path
+ *         required: true
+ *         schema:
+ *           type: string
+ *         description: Note ID to get OCR text for
+ *     responses:
+ *       200:
+ *         description: OCR text retrieved successfully
+ *       404:
+ *         description: Note not found
+ *     tags: ["ocr"]
+ */
+function getTextRepresentation(blobId: string | undefined): TextRepresentationResponse {
+    let ocrText: string | null = null;
+
+    if (blobId) {
+        const result = sql.getRow<{
+            textRepresentation: string | null;
+        }>(`
+            SELECT textRepresentation
+            FROM blobs
+            WHERE blobId = ?
+        `, [blobId]);
+
+        if (result) {
+            ocrText = result.textRepresentation;
+        }
+    }
+
+    return {
+        success: true,
+        text: ocrText || '',
+        hasOcr: !!ocrText
+    };
+}
+
+async function getNoteOCRText(req: Request<{ noteId: string }>) {
+    const note = becca.getNote(req.params.noteId);
+    if (!note) {
+        return [404, { success: false, message: 'Note not found' }];
+    }
+
+    return getTextRepresentation(note.blobId);
+}
+
+async function getAttachmentOCRText(req: Request<{ attachmentId: string }>) {
+    const attachment = becca.getAttachment(req.params.attachmentId);
+    if (!attachment) {
+        return [404, { success: false, message: 'Attachment not found' }];
+    }
+
+    return getTextRepresentation(attachment.blobId);
+}
+
+export default {
+    processNoteOCR,
+    processAttachmentOCR,
+    batchProcessOCR,
+    getBatchProgress,
+    getNoteOCRText,
+    getAttachmentOCRText
+};
--- a/apps/server/src/routes/api/options.ts
+++ b/apps/server/src/routes/api/options.ts
@@ -105,7 +105,11 @@ const ALLOWED_OPTIONS = new Set<OptionNames>([
    "newLayout",
    "mfaEnabled",
    "mfaMethod",
-    "llmProviders"
+    "llmProviders",
+
+    // OCR options
+    "ocrAutoProcessImages",
+    "ocrMinConfidence"
 ]);

 function getOptions() {
--- a/apps/server/src/routes/routes.ts
+++ b/apps/server/src/routes/routes.ts
@@ -39,6 +39,7 @@ import loginApiRoute from "./api/login.js";
 import metricsRoute from "./api/metrics.js";
 import noteMapRoute from "./api/note_map.js";
 import notesApiRoute from "./api/notes.js";
+import ocrRoute from "./api/ocr.js";
 import optionsApiRoute from "./api/options.js";
 import otherRoute from "./api/other.js";
 import passwordApiRoute from "./api/password.js";
@@ -376,6 +377,14 @@ function register(app: express.Application) {
    etapiBackupRoute.register(router);
    etapiMetricsRoute.register(router);

+    // OCR API
+    asyncApiRoute(PST, "/api/ocr/process-note/:noteId", ocrRoute.processNoteOCR);
+    asyncApiRoute(PST, "/api/ocr/process-attachment/:attachmentId", ocrRoute.processAttachmentOCR);
+    asyncApiRoute(PST, "/api/ocr/batch-process", ocrRoute.batchProcessOCR);
+    asyncApiRoute(GET, "/api/ocr/batch-progress", ocrRoute.getBatchProgress);
+    asyncApiRoute(GET, "/api/ocr/notes/:noteId/text", ocrRoute.getNoteOCRText);
+    asyncApiRoute(GET, "/api/ocr/attachments/:attachmentId/text", ocrRoute.getAttachmentOCRText);
+
    app.use("", router);
 }

--- a/apps/server/src/services/app_info.ts
+++ b/apps/server/src/services/app_info.ts
@@ -5,7 +5,7 @@ import packageJson from "../../package.json" with { type: "json" };
 import build from "./build.js";
 import dataDir from "./data_dir.js";

-const APP_DB_VERSION = 235;
+const APP_DB_VERSION = 236;
 const SYNC_VERSION = 37;
 const CLIPPER_PROTOCOL_VERSION = "1.0";

--- a/apps/server/src/services/blob-interface.ts
+++ b/apps/server/src/services/blob-interface.ts
@@ -1,5 +1,6 @@
 export interface Blob {
    blobId: string;
    content: string | Buffer;
+    textRepresentation?: string | null;
    utcDateModified: string;
 }
--- a/apps/server/src/services/blob.ts
+++ b/apps/server/src/services/blob.ts
@@ -50,8 +50,8 @@ function processContent(content: Buffer | string | null, isProtected: boolean, i
    }
 }

-function calculateContentHash({ blobId, content }: Blob) {
-    return hash(`${blobId}|${content.toString()}`);
+function calculateContentHash({ blobId, content, textRepresentation }: Blob) {
+    return hash(`${blobId}|${content.toString()}|${textRepresentation ?? ""}`);
 }

 export default {
--- a/apps/server/src/services/data_dir.spec.ts
+++ b/apps/server/src/services/data_dir.spec.ts
@@ -1,6 +1,6 @@
-import { describe, it, expect, beforeEach, vi } from "vitest";
+import { beforeEach, describe, expect, it, vi } from "vitest";

-import type { getTriliumDataDir as getTriliumDataDirType, getDataDirs as getDataDirsType, getPlatformAppDataDir as getPlatformAppDataDirType } from "./data_dir.js";
+import type { getDataDirs as getDataDirsType, getPlatformAppDataDir as getPlatformAppDataDirType,getTriliumDataDir as getTriliumDataDirType } from "./data_dir.js";

 describe("data_dir.ts unit tests", async () => {
    let getTriliumDataDir: typeof getTriliumDataDirType;
@@ -277,7 +277,7 @@ describe("data_dir.ts unit tests", async () => {
    });

    describe("#getDataDirs()", () => {
-        const envKeys: Omit<keyof ReturnType<typeof getDataDirs>, "TRILIUM_DATA_DIR">[] = [ "DOCUMENT_PATH", "BACKUP_DIR", "LOG_DIR", "ANONYMIZED_DB_DIR", "CONFIG_INI_PATH", "TMP_DIR" ];
+        const envKeys: Omit<keyof ReturnType<typeof getDataDirs>, "TRILIUM_DATA_DIR">[] = [ "DOCUMENT_PATH", "BACKUP_DIR", "LOG_DIR", "ANONYMIZED_DB_DIR", "CONFIG_INI_PATH", "TMP_DIR", "OCR_CACHE_DIR" ];

        const setMockedEnv = (prefix: string | null) => {
            envKeys.forEach((key) => {
--- a/apps/server/src/services/data_dir.ts
+++ b/apps/server/src/services/data_dir.ts
@@ -1,5 +1,3 @@
-"use strict";
-
 /*
 * This file resolves trilium data path in this order of priority:
 * - case A) if TRILIUM_DATA_DIR environment variable exists, then its value is used as the path
@@ -8,8 +6,8 @@
 * - case D) as a fallback if the previous step fails, we'll use home dir
 */

-import os from "os";
 import fs from "fs";
+import os from "os";
 import { join as pathJoin } from "path";

 const DIR_NAME = "trilium-data";
@@ -43,13 +41,14 @@ export function getTriliumDataDir(dataDirName: string) {

 export function getDataDirs(TRILIUM_DATA_DIR: string) {
    const dataDirs = {
-        TRILIUM_DATA_DIR: TRILIUM_DATA_DIR,
+        TRILIUM_DATA_DIR,
        DOCUMENT_PATH: process.env.TRILIUM_DOCUMENT_PATH || pathJoin(TRILIUM_DATA_DIR, "document.db"),
        BACKUP_DIR: process.env.TRILIUM_BACKUP_DIR || pathJoin(TRILIUM_DATA_DIR, "backup"),
        LOG_DIR: process.env.TRILIUM_LOG_DIR || pathJoin(TRILIUM_DATA_DIR, "log"),
        TMP_DIR: process.env.TRILIUM_TMP_DIR || pathJoin(TRILIUM_DATA_DIR, "tmp"),
        ANONYMIZED_DB_DIR: process.env.TRILIUM_ANONYMIZED_DB_DIR || pathJoin(TRILIUM_DATA_DIR, "anonymized-db"),
-        CONFIG_INI_PATH: process.env.TRILIUM_CONFIG_INI_PATH || pathJoin(TRILIUM_DATA_DIR, "config.ini")
+        CONFIG_INI_PATH: process.env.TRILIUM_CONFIG_INI_PATH || pathJoin(TRILIUM_DATA_DIR, "config.ini"),
+        OCR_CACHE_DIR: process.env.TRILIUM_OCR_CACHE_DIR || pathJoin(TRILIUM_DATA_DIR, "ocr-cache")
    } as const;

    createDirIfNotExisting(dataDirs.TMP_DIR);
--- a/apps/server/src/services/entity_changes.ts
+++ b/apps/server/src/services/entity_changes.ts
@@ -146,7 +146,7 @@ function fillEntityChanges(entityName: string, entityPrimaryKey: string, conditi
            };

            if (entityName === "blobs") {
-                const blob = sql.getRow<Blob>("SELECT blobId, content, utcDateModified FROM blobs WHERE blobId = ?", [entityId]);
+                const blob = sql.getRow<Blob>("SELECT blobId, content, textRepresentation, utcDateModified FROM blobs WHERE blobId = ?", [entityId]);
                ec.hash = blobService.calculateContentHash(blob);
                ec.utcDateChanged = blob.utcDateModified;
                ec.isSynced = true; // blobs are always synced
--- a/apps/server/src/services/handlers.ts
+++ b/apps/server/src/services/handlers.ts
@@ -6,6 +6,9 @@ import becca from "../becca/becca.js";
 import BAttribute from "../becca/entities/battribute.js";
 import hiddenSubtreeService from "./hidden_subtree.js";
 import oneTimeTimer from "./one_time_timer.js";
+import ocrService from "./ocr/ocr_service.js";
+import optionService from "./options.js";
+import log from "./log.js";
 import type BNote from "../becca/entities/bnote.js";
 import type AbstractBeccaEntity from "../becca/entities/abstract_becca_entity.js";
 import type { DefinitionObject } from "./promoted_attribute_definition_interface.js";
@@ -137,9 +140,35 @@ eventService.subscribe(eventService.ENTITY_CREATED, ({ entityName, entity }) =>
        }
    } else if (entityName === "notes") {
        runAttachedRelations(entity, "runOnNoteCreation", entity);
+
+        // Note: OCR processing for images is now handled in image.ts during image processing
+        // OCR processing for files remains here since they don't go through image processing
+        if (entity.type === 'file' && optionService.getOptionBool("ocrAutoProcessImages")) {
+            autoProcessOCR(entity.mime, () => ocrService.processNoteOCR(entity.noteId), `file note ${entity.noteId}`);
+        }
+    } else if (entityName === "attachments") {
+        // Image attachments are handled in image.ts after async image processing sets the real MIME type.
+        // Only handle non-image (file) attachments here.
+        if (entity.role === "file" && optionService.getOptionBool("ocrAutoProcessImages")) {
+            autoProcessOCR(entity.mime, () => ocrService.processAttachmentOCR(entity.attachmentId), `attachment ${entity.attachmentId}`);
+        }
    }
 });

+function autoProcessOCR(mime: string, process: () => Promise<unknown>, entityDescription: string) {
+    const supportedMimeTypes = ocrService.getAllSupportedMimeTypes();
+
+    if (mime && supportedMimeTypes.includes(mime)) {
+        process().then(result => {
+            if (result) {
+                log.info(`Automatically processed OCR for ${entityDescription} with MIME type ${mime}`);
+            }
+        }).catch(error => {
+            log.error(`Failed to automatically process OCR for ${entityDescription}: ${error}`);
+        });
+    }
+}
+
 eventService.subscribe(eventService.CHILD_NOTE_CREATED, ({ parentNote, childNote }) => {
    runAttachedRelations(parentNote, "runOnChildNoteCreation", childNote);
 });
--- a/apps/server/src/services/hidden_subtree.ts
+++ b/apps/server/src/services/hidden_subtree.ts
@@ -246,7 +246,8 @@ function buildHiddenSubtreeDefinition(helpSubtree: HiddenSubtreeItem[]): HiddenS
                    { id: "_optionsShortcuts", title: t("hidden-subtree.shortcuts-title"), type: "contentWidget", icon: "bxs-keyboard" },
                    { id: "_optionsTextNotes", title: t("hidden-subtree.text-notes"), type: "contentWidget", icon: "bx-text" },
                    { id: "_optionsCodeNotes", title: t("hidden-subtree.code-notes-title"), type: "contentWidget", icon: "bx-code" },
-                    { id: "_optionsImages", title: t("hidden-subtree.images-title"), type: "contentWidget", icon: "bx-image" },
+                    { id: "_optionsImages", title: "Images", type: "contentWidget", enforceDeleted: true },
+                    { id: "_optionsMedia", title: t("hidden-subtree.images-title"), type: "contentWidget", icon: "bx-image" },
                    { id: "_optionsSpellcheck", title: t("hidden-subtree.spellcheck-title"), type: "contentWidget", icon: "bx-check-double" },
                    { id: "_optionsPassword", title: t("hidden-subtree.password-title"), type: "contentWidget", icon: "bx-lock" },
                    { id: '_optionsMFA', title: t('hidden-subtree.multi-factor-authentication-title'), type: 'contentWidget', icon: 'bx-lock ' },
--- a/apps/server/src/services/image.ts
+++ b/apps/server/src/services/image.ts
@@ -1,17 +1,17 @@
-"use strict";
+import imageType from "image-type";
+import isAnimated from "is-animated";
+import isSvg from "is-svg";
+import { Jimp } from "jimp";
+import sanitizeFilename from "sanitize-filename";

 import becca from "../becca/becca.js";
-import log from "./log.js";
-import protectedSessionService from "./protected_session.js";
-import noteService from "./notes.js";
-import optionService from "./options.js";
-import sql from "./sql.js";
-import { Jimp } from "jimp";
-import imageType from "image-type";
-import sanitizeFilename from "sanitize-filename";
-import isSvg from "is-svg";
-import isAnimated from "is-animated";
 import htmlSanitizer from "./html_sanitizer.js";
+import log from "./log.js";
+import noteService from "./notes.js";
+import ocrService from "./ocr/ocr_service.js";
+import optionService from "./options.js";
+import protectedSessionService from "./protected_session.js";
+import sql from "./sql.js";

 async function processImage(uploadBuffer: Buffer, originalName: string, shrinkImageSwitch: boolean) {
    const compressImages = optionService.getOptionBool("compressImages");
@@ -46,9 +46,8 @@ async function processImage(uploadBuffer: Buffer, originalName: string, shrinkIm
 async function getImageType(buffer: Buffer) {
    if (isSvg(buffer.toString())) {
        return { ext: "svg" };
-    } else {
-        return (await imageType(buffer)) || { ext: "jpg" }; // optimistic JPG default
    }
+    return (await imageType(buffer)) || { ext: "jpg" }; // optimistic JPG default
 }

 function getImageMimeFromExtension(ext: string) {
@@ -79,6 +78,8 @@ function updateImage(noteId: string, uploadBuffer: Buffer, originalName: string)

            note.setContent(buffer);
        });
+
+        scheduleOcrForNote(noteId);
    });
 }

@@ -121,6 +122,8 @@ function saveImage(parentNoteId: string, uploadBuffer: Buffer, originalName: str

            note.setContent(buffer, { forceSave: true });
        });
+
+        scheduleOcrForNote(note.noteId);
    });

    return {
@@ -159,13 +162,14 @@ function saveImageToAttachment(noteId: string, uploadBuffer: Buffer, originalNam
    }, 5000);

    // resizing images asynchronously since JIMP does not support sync operation
+    const attachmentId = attachment.attachmentId;
    processImage(uploadBuffer, originalName, !!shrinkImageSwitch).then(({ buffer, imageFormat }) => {
        sql.transactional(() => {
            // re-read, might be changed in the meantime
-            if (!attachment.attachmentId) {
+            if (!attachmentId) {
                throw new Error("Missing attachment ID.");
            }
-            attachment = becca.getAttachmentOrThrow(attachment.attachmentId);
+            attachment = becca.getAttachmentOrThrow(attachmentId);

            attachment.mime = getImageMimeFromExtension(imageFormat.ext);

@@ -176,11 +180,37 @@ function saveImageToAttachment(noteId: string, uploadBuffer: Buffer, originalNam

            attachment.setContent(buffer, { forceSave: true });
        });
+
+        scheduleOcrForAttachment(attachmentId);
    });

    return attachment;
 }

+function scheduleOcrForNote(noteId: string) {
+    if (optionService.getOptionBool("ocrAutoProcessImages")) {
+        setImmediate(async () => {
+            try {
+                await ocrService.processNoteOCR(noteId);
+            } catch (error) {
+                log.error(`Failed to process OCR for note ${noteId}: ${error}`);
+            }
+        });
+    }
+}
+
+function scheduleOcrForAttachment(attachmentId: string | undefined) {
+    if (attachmentId && optionService.getOptionBool("ocrAutoProcessImages")) {
+        setImmediate(async () => {
+            try {
+                await ocrService.processAttachmentOCR(attachmentId);
+            } catch (error) {
+                log.error(`Failed to process OCR for attachment ${attachmentId}: ${error}`);
+            }
+        });
+    }
+}
+
 async function shrinkImage(buffer: Buffer, originalName: string) {
    let jpegQuality = optionService.getOptionInt("imageJpegQuality", 0);

--- a/apps/server/src/services/llm/tools/note_tools.ts
+++ b/apps/server/src/services/llm/tools/note_tools.ts
@@ -16,9 +16,14 @@ import searchService from "../../search/services/search.js";
 * Convert note content to a format suitable for LLM consumption.
 * Text notes are converted from HTML to Markdown to reduce token usage.
 */
-export function getNoteContentForLlm(note: { type: string; getContent: () => string | Buffer }) {
+export function getNoteContentForLlm(note: { type: string; blobId?: string; getContent: () => string | Buffer }) {
    const content = note.getContent();
    if (typeof content !== "string") {
+        // For binary content (images, files), use extracted text if available.
+        const blob = note.blobId ? becca.getBlob({ blobId: note.blobId }) : null;
+        if (blob?.textRepresentation) {
+            return `[extracted text from ${note.type}]\n${blob.textRepresentation}`;
+        }
        return "[binary content]";
    }
    if (note.type === "text") {
--- a/apps/server/src/services/ocr/ocr_service.spec.ts
+++ b/apps/server/src/services/ocr/ocr_service.spec.ts
@@ -0,0 +1,450 @@
+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
+
+// Mock Tesseract.js
+const mockWorker = {
+    recognize: vi.fn(),
+    terminate: vi.fn(),
+    reinitialize: vi.fn()
+};
+
+const mockTesseract = {
+    createWorker: vi.fn().mockResolvedValue(mockWorker)
+};
+
+vi.mock('tesseract.js', () => ({
+    default: mockTesseract
+}));
+
+// Mock dependencies
+const mockOptions = {
+    getOptionBool: vi.fn(),
+    getOption: vi.fn()
+};
+
+const mockLog = {
+    info: vi.fn(),
+    error: vi.fn()
+};
+
+const mockSql = {
+    execute: vi.fn(),
+    getRow: vi.fn(),
+    getRows: vi.fn(),
+    getColumn: vi.fn()
+};
+
+const mockBecca = {
+    getNote: vi.fn(),
+    getAttachment: vi.fn(),
+    getBlob: vi.fn()
+};
+
+const mockBlobService = {
+    calculateContentHash: vi.fn().mockReturnValue('hash123')
+};
+
+const mockEntityChangesService = {
+    putEntityChange: vi.fn()
+};
+
+vi.mock('../options.js', () => ({
+    default: mockOptions
+}));
+
+vi.mock('../log.js', () => ({
+    default: mockLog
+}));
+
+vi.mock('../sql.js', () => ({
+    default: mockSql
+}));
+
+vi.mock('../../becca/becca.js', () => ({
+    default: mockBecca
+}));
+
+vi.mock('../blob.js', () => ({
+    default: mockBlobService
+}));
+
+vi.mock('../entity_changes.js', () => ({
+    default: mockEntityChangesService
+}));
+
+// Import the service after mocking
+let ocrService: typeof import('./ocr_service.js').default;
+
+beforeEach(async () => {
+    vi.clearAllMocks();
+
+    // Reset mock implementations
+    mockOptions.getOptionBool.mockReturnValue(true);
+    mockOptions.getOption.mockImplementation((name: string) => {
+        if (name === 'ocrMinConfidence') return '0';
+        return 'eng';
+    });
+    mockSql.execute.mockImplementation(() => ({ lastInsertRowid: 1 }));
+    mockSql.getRow.mockReturnValue(null);
+    mockSql.getRows.mockReturnValue([]);
+    mockSql.getColumn.mockReturnValue([]);
+
+    // Mock getBlob for putBlobEntityChange
+    mockBecca.getBlob.mockReturnValue({
+        blobId: 'blob123',
+        content: Buffer.from('data'),
+        textRepresentation: null,
+        utcDateModified: '2025-01-01'
+    });
+
+    mockTesseract.createWorker.mockImplementation(async () => {
+        return mockWorker;
+    });
+
+    // Dynamically import the service to ensure mocks are applied
+    const module = await import('./ocr_service.js');
+    ocrService = module.default;
+
+    // Reset the OCR service state
+    (ocrService as any).batchProcessingState = {
+        inProgress: false,
+        total: 0,
+        processed: 0
+    };
+});
+
+afterEach(() => {
+    vi.restoreAllMocks();
+});
+
+describe('OCRService', () => {
+    describe('extractTextFromFile', () => {
+        const mockImageBuffer = Buffer.from('fake-image-data');
+
+        it('should extract text successfully with default options', async () => {
+            const mockResult = {
+                data: {
+                    text: 'Extracted text from image',
+                    confidence: 95,
+                    words: [{ text: 'Extracted', confidence: 95 }, { text: 'text', confidence: 95 }, { text: 'from', confidence: 95 }, { text: 'image', confidence: 95 }]
+                }
+            };
+            mockWorker.recognize.mockResolvedValue(mockResult);
+
+            const result = await ocrService.extractTextFromFile(mockImageBuffer, 'image/jpeg');
+
+            expect(result).toBeDefined();
+            expect(result.text).toBe('Extracted text from image');
+            expect(result.extractedAt).toEqual(expect.any(String));
+        });
+
+        it('should handle OCR recognition errors', async () => {
+            const error = new Error('OCR recognition failed');
+            mockWorker.recognize.mockRejectedValue(error);
+
+            await expect(ocrService.extractTextFromFile(mockImageBuffer, 'image/jpeg')).rejects.toThrow('OCR recognition failed');
+            expect(mockLog.error).toHaveBeenCalledWith('Image OCR text extraction failed: Error: OCR recognition failed');
+        });
+    });
+
+    describe('storeOCRResult', () => {
+        it('should store OCR result in blob successfully', () => {
+            const ocrResult = {
+                text: 'Sample text',
+                confidence: 0.95,
+                extractedAt: '2025-06-10T10:00:00.000Z',
+                language: 'eng'
+            };
+
+            ocrService.storeOCRResult('blob123', ocrResult);
+
+            expect(mockSql.execute).toHaveBeenCalledWith(
+                expect.stringContaining('UPDATE blobs SET textRepresentation = ?'),
+                ['Sample text', 'blob123']
+            );
+        });
+
+        it('should handle undefined blobId gracefully', () => {
+            const ocrResult = {
+                text: 'Sample text',
+                confidence: 0.95,
+                extractedAt: '2025-06-10T10:00:00.000Z',
+                language: 'eng'
+            };
+
+            ocrService.storeOCRResult(undefined, ocrResult);
+
+            expect(mockSql.execute).not.toHaveBeenCalled();
+            expect(mockLog.error).toHaveBeenCalledWith('Cannot store OCR result: blobId is undefined');
+        });
+
+        it('should handle database update errors', () => {
+            const error = new Error('Database error');
+            mockSql.execute.mockImplementation(() => {
+                throw error;
+            });
+
+            const ocrResult = {
+                text: 'Sample text',
+                confidence: 0.95,
+                extractedAt: '2025-06-10T10:00:00.000Z',
+                language: 'eng'
+            };
+
+            expect(() => ocrService.storeOCRResult('blob123', ocrResult)).toThrow('Database error');
+            expect(mockLog.error).toHaveBeenCalledWith('Failed to store OCR result for blob blob123: Error: Database error');
+        });
+    });
+
+    describe('processNoteOCR', () => {
+        const mockNote = {
+            noteId: 'note123',
+            type: 'image',
+            mime: 'image/jpeg',
+            blobId: 'blob123',
+            getContent: vi.fn(),
+            getLabelValue: vi.fn().mockReturnValue(null)
+        };
+
+        beforeEach(() => {
+            mockBecca.getNote.mockReturnValue(mockNote);
+            mockNote.getContent.mockReturnValue(Buffer.from('fake-image-data'));
+            mockNote.mime = 'image/jpeg';
+        });
+
+        it('should process note OCR successfully', async () => {
+            mockSql.getRow.mockReturnValue(null);
+
+            const mockOCRResult = {
+                data: {
+                    text: 'Note image text',
+                    confidence: 90,
+                    words: [{ text: 'Note', confidence: 90 }, { text: 'image', confidence: 90 }, { text: 'text', confidence: 90 }]
+                }
+            };
+            mockWorker.recognize.mockResolvedValue(mockOCRResult);
+
+            const result = await ocrService.processNoteOCR('note123');
+
+            expect(result).toBeDefined();
+            expect(result!.text).toBe('Note image text');
+            expect(mockBecca.getNote).toHaveBeenCalledWith('note123');
+            expect(mockNote.getContent).toHaveBeenCalled();
+        });
+
+        it('should skip processing if OCR already exists and forceReprocess is false', async () => {
+            mockSql.getRow.mockReturnValue({ textRepresentation: 'Existing text' });
+
+            const result = await ocrService.processNoteOCR('note123');
+
+            expect(result).toBeNull();
+            expect(mockNote.getContent).not.toHaveBeenCalled();
+        });
+
+        it('should reprocess if forceReprocess is true', async () => {
+            mockSql.getRow.mockReturnValue({ textRepresentation: 'Existing text' });
+
+            const mockOCRResult = {
+                data: {
+                    text: 'New processed text',
+                    confidence: 95,
+                    words: [{ text: 'New', confidence: 95 }, { text: 'processed', confidence: 95 }, { text: 'text', confidence: 95 }]
+                }
+            };
+            mockWorker.recognize.mockResolvedValue(mockOCRResult);
+
+            const result = await ocrService.processNoteOCR('note123', { forceReprocess: true });
+
+            expect(result?.text).toBe('New processed text');
+            expect(mockNote.getContent).toHaveBeenCalled();
+        });
+
+        it('should return null for non-existent note', async () => {
+            mockBecca.getNote.mockReturnValue(null);
+
+            const result = await ocrService.processNoteOCR('nonexistent');
+
+            expect(result).toBe(null);
+            expect(mockLog.error).toHaveBeenCalledWith('Note nonexistent not found');
+        });
+
+        it('should return null for unsupported MIME type', async () => {
+            mockNote.mime = 'text/plain';
+
+            const result = await ocrService.processNoteOCR('note123');
+
+            expect(result).toBe(null);
+            expect(mockLog.info).toHaveBeenCalledWith('note note123 has unsupported MIME type text/plain for text extraction, skipping');
+        });
+    });
+
+    describe('processAttachmentOCR', () => {
+        const mockAttachment = {
+            attachmentId: 'attach123',
+            ownerId: 'note123',
+            role: 'image',
+            mime: 'image/png',
+            blobId: 'blob456',
+            getContent: vi.fn()
+        };
+
+        beforeEach(() => {
+            mockBecca.getAttachment.mockReturnValue(mockAttachment);
+            mockBecca.getNote.mockReturnValue({ getLabelValue: vi.fn().mockReturnValue(null) });
+            mockAttachment.getContent.mockReturnValue(Buffer.from('fake-image-data'));
+        });
+
+        it('should process attachment OCR successfully', async () => {
+            mockSql.getRow.mockReturnValue(null);
+
+            const mockOCRResult = {
+                data: {
+                    text: 'Attachment image text',
+                    confidence: 92,
+                    words: [{ text: 'Attachment', confidence: 92 }, { text: 'image', confidence: 92 }, { text: 'text', confidence: 92 }]
+                }
+            };
+            mockWorker.recognize.mockResolvedValue(mockOCRResult);
+
+            const result = await ocrService.processAttachmentOCR('attach123');
+
+            expect(result).toBeDefined();
+            expect(result!.text).toBe('Attachment image text');
+            expect(mockBecca.getAttachment).toHaveBeenCalledWith('attach123');
+        });
+
+        it('should return null for non-existent attachment', async () => {
+            mockBecca.getAttachment.mockReturnValue(null);
+
+            const result = await ocrService.processAttachmentOCR('nonexistent');
+
+            expect(result).toBe(null);
+            expect(mockLog.error).toHaveBeenCalledWith('Attachment nonexistent not found');
+        });
+    });
+
+    describe('Batch Processing', () => {
+        // Helper to mock getBlobsNeedingOCR to return entities
+        function mockBlobsNeedingOCR(notes: Array<{ entityId: string; mimeType: string }>, attachments: Array<{ entityId: string; mimeType: string }> = []) {
+            const noteRows = notes.map(n => ({ blobId: `blob_${n.entityId}`, mimeType: n.mimeType, entityId: n.entityId }));
+            const attachmentRows = attachments.map(a => ({ blobId: `blob_${a.entityId}`, mimeType: a.mimeType, entityId: a.entityId }));
+            mockSql.getRows.mockReturnValueOnce(noteRows);
+            mockSql.getRows.mockReturnValueOnce(attachmentRows);
+        }
+
+        describe('startBatchProcessing', () => {
+            beforeEach(() => {
+                ocrService.cancelBatchProcessing();
+            });
+
+            it('should start batch processing when items are available', async () => {
+                mockBlobsNeedingOCR(
+                    [{ entityId: 'note1', mimeType: 'image/jpeg' }]
+                );
+
+                const result = await ocrService.startBatchProcessing();
+
+                expect(result).toEqual({ success: true });
+            });
+
+            it('should return error if batch processing already in progress', async () => {
+                // First call: items for starting
+                mockBlobsNeedingOCR(
+                    [{ entityId: 'note1', mimeType: 'image/jpeg' }]
+                );
+                // Mock note for background processing
+                mockBecca.getNote.mockReturnValue({
+                    noteId: 'note1', type: 'image', mime: 'image/jpeg', blobId: 'blob1',
+                    getContent: vi.fn().mockReturnValue(Buffer.from('data')),
+                    getLabelValue: vi.fn().mockReturnValue(null)
+                });
+                mockWorker.recognize.mockResolvedValue({ data: { text: 'text', confidence: 90, words: [] } });
+
+                ocrService.startBatchProcessing();
+
+                const result = await ocrService.startBatchProcessing();
+
+                expect(result).toEqual({
+                    success: false,
+                    message: 'Batch processing already in progress'
+                });
+            });
+
+            it('should return error if no items need processing', async () => {
+                mockBlobsNeedingOCR([], []);
+
+                const result = await ocrService.startBatchProcessing();
+
+                expect(result).toEqual({
+                    success: false,
+                    message: 'No images found that need OCR processing'
+                });
+            });
+
+            it('should handle database errors gracefully', async () => {
+                mockSql.getRows.mockImplementation(() => {
+                    throw new Error('Database connection failed');
+                });
+
+                const result = await ocrService.startBatchProcessing();
+
+                // getBlobsNeedingOCR catches DB errors and returns [], so startBatchProcessing sees no items
+                expect(result).toEqual({
+                    success: false,
+                    message: 'No images found that need OCR processing'
+                });
+                expect(mockLog.error).toHaveBeenCalledWith(
+                    expect.stringContaining('Failed to get blobs needing OCR')
+                );
+            });
+        });
+
+        describe('getBatchProgress', () => {
+            it('should return initial progress state', () => {
+                const progress = ocrService.getBatchProgress();
+
+                expect(progress.inProgress).toBe(false);
+                expect(progress.total).toBe(0);
+                expect(progress.processed).toBe(0);
+            });
+
+            it('should return progress with percentage when total > 0', async () => {
+                mockBlobsNeedingOCR(
+                    Array.from({ length: 10 }, (_, i) => ({ entityId: `note${i}`, mimeType: 'image/jpeg' }))
+                );
+
+                ocrService.startBatchProcessing();
+
+                const progress = ocrService.getBatchProgress();
+
+                expect(progress.inProgress).toBe(true);
+                expect(progress.total).toBe(10);
+                expect(progress.processed).toBe(0);
+                expect(progress.percentage).toBe(0);
+                expect(progress.startTime).toBeInstanceOf(Date);
+            });
+        });
+
+        describe('cancelBatchProcessing', () => {
+            it('should cancel ongoing batch processing', async () => {
+                mockBlobsNeedingOCR(
+                    [{ entityId: 'note1', mimeType: 'image/jpeg' }]
+                );
+
+                ocrService.startBatchProcessing();
+
+                expect(ocrService.getBatchProgress().inProgress).toBe(true);
+
+                ocrService.cancelBatchProcessing();
+
+                expect(ocrService.getBatchProgress().inProgress).toBe(false);
+                expect(mockLog.info).toHaveBeenCalledWith('Batch OCR processing cancelled');
+            });
+
+            it('should do nothing if no batch processing is running', () => {
+                ocrService.cancelBatchProcessing();
+
+                expect(mockLog.info).not.toHaveBeenCalledWith('Batch OCR processing cancelled');
+            });
+        });
+    });
+});
--- a/apps/server/src/services/ocr/ocr_service.ts
+++ b/apps/server/src/services/ocr/ocr_service.ts
@@ -0,0 +1,462 @@
+import { getTesseractCode } from '@triliumnext/commons';
+
+import becca from '../../becca/becca.js';
+import blobService from '../blob.js';
+import entityChangesService from '../entity_changes.js';
+import log from '../log.js';
+import options from '../options.js';
+import sql from '../sql.js';
+import { FileProcessor } from './processors/file_processor.js';
+import { ImageProcessor } from './processors/image_processor.js';
+import { OfficeProcessor } from './processors/office_processor.js';
+import { PDFProcessor } from './processors/pdf_processor.js';
+
+export interface OCRResult {
+    text: string;
+    confidence: number;
+    extractedAt: string;
+    language?: string;
+    pageCount?: number;
+}
+
+export interface OCRProcessingOptions {
+    language?: string;
+    forceReprocess?: boolean;
+    confidence?: number;
+    enablePDFTextExtraction?: boolean;
+    mimeType?: string;
+}
+
+/**
+ * OCR Service for extracting text from images and other OCR-able objects
+ * Uses Tesseract.js for text recognition
+ */
+class OCRService {
+    private processors: Map<string, FileProcessor> = new Map();
+
+    constructor() {
+        this.processors.set('image', new ImageProcessor());
+        this.processors.set('pdf', new PDFProcessor());
+        this.processors.set('office', new OfficeProcessor());
+    }
+
+    /**
+     * Resolves the Tesseract language code(s) for OCR processing.
+     *
+     * Priority:
+     * 1. Explicitly passed `language` option (e.g. from API call)
+     * 2. The note's `language` label (mapped via {@link getTesseractCode})
+     * 3. All enabled content languages joined with `+`
+     * 4. The UI locale
+     * 5. Fallback to `eng`
+     */
+    resolveOcrLanguage(noteId?: string, explicitLanguage?: string): string {
+        // 1. Explicit language from caller
+        if (explicitLanguage) {
+            return explicitLanguage;
+        }
+
+        // 2. Note's language label
+        if (noteId) {
+            const note = becca.getNote(noteId);
+            const noteLanguage = note?.getLabelValue("language");
+            if (noteLanguage) {
+                const code = getTesseractCode(noteLanguage);
+                if (code) {
+                    return code;
+                }
+            }
+        }
+
+        // 3. All enabled content languages
+        try {
+            const languagesJson = options.getOption("languages");
+            const enabledLanguages = JSON.parse(languagesJson || "[]") as string[];
+            if (enabledLanguages.length > 0) {
+                const codes = enabledLanguages
+                    .map((id) => getTesseractCode(id))
+                    .filter((code): code is string => code !== null);
+                // Deduplicate (e.g. en + en-GB both map to eng)
+                const unique = [...new Set(codes)];
+                if (unique.length > 0) {
+                    return unique.join("+");
+                }
+            }
+        } catch {
+            // Fall through
+        }
+
+        // 4. UI locale
+        try {
+            const uiLocale = options.getOption("locale");
+            if (uiLocale) {
+                const code = getTesseractCode(uiLocale);
+                if (code) {
+                    return code;
+                }
+            }
+        } catch {
+            // Fall through
+        }
+
+        // 5. Fallback
+        return "eng";
+    }
+
+
+    /**
+     * Extract text from file buffer using appropriate processor
+     */
+    async extractTextFromFile(fileBuffer: Buffer, mimeType: string, options: OCRProcessingOptions = {}): Promise<OCRResult> {
+        log.info(`Starting OCR text extraction for MIME type: ${mimeType} with language: ${options.language || "eng"}`);
+
+        const processor = this.getProcessorForMimeType(mimeType);
+        if (!processor) {
+            throw new Error(`No processor found for MIME type: ${mimeType}`);
+        }
+
+        const result = await processor.extractText(fileBuffer, { ...options, mimeType });
+
+        log.info(`OCR extraction completed. Confidence: ${Math.round(result.confidence * 100)}%, Text length: ${result.text.length}`);
+        return result;
+    }
+
+    /**
+     * Process OCR for a note (image type)
+     */
+    async processNoteOCR(noteId: string, options: OCRProcessingOptions = {}): Promise<OCRResult | null> {
+        const note = becca.getNote(noteId);
+        if (!note) {
+            log.error(`Note ${noteId} not found`);
+            return null;
+        }
+
+        return this.processEntityOCR({
+            entityId: noteId,
+            entityType: 'note',
+            category: note.type,
+            mime: note.mime,
+            blobId: note.blobId,
+            languageNoteId: noteId,
+            getContent: () => note.getContent()
+        }, options);
+    }
+
+    /**
+     * Process OCR for an attachment
+     */
+    async processAttachmentOCR(attachmentId: string, options: OCRProcessingOptions = {}): Promise<OCRResult | null> {
+        const attachment = becca.getAttachment(attachmentId);
+        if (!attachment) {
+            log.error(`Attachment ${attachmentId} not found`);
+            return null;
+        }
+
+        return this.processEntityOCR({
+            entityId: attachmentId,
+            entityType: 'attachment',
+            category: attachment.role,
+            mime: attachment.mime,
+            blobId: attachment.blobId,
+            languageNoteId: attachment.ownerId,
+            getContent: () => attachment.getContent()
+        }, options);
+    }
+
+    /**
+     * Shared OCR processing logic for both notes and attachments.
+     */
+    private async processEntityOCR(entity: {
+        entityId: string;
+        entityType: string;
+        category: string;
+        mime: string;
+        blobId: string | undefined;
+        languageNoteId: string;
+        getContent: () => string | Buffer;
+    }, options: OCRProcessingOptions = {}): Promise<OCRResult | null> {
+        const { entityId, entityType, category, mime, blobId, languageNoteId } = entity;
+
+        if (!['image', 'file'].includes(category)) {
+            log.info(`${entityType} ${entityId} is not an image or file, skipping OCR`);
+            return null;
+        }
+
+        if (!this.getProcessorForMimeType(mime)) {
+            log.info(`${entityType} ${entityId} has unsupported MIME type ${mime} for text extraction, skipping`);
+            return null;
+        }
+
+        if (!options.forceReprocess && this.hasStoredOCRResult(blobId)) {
+            log.info(`OCR already exists for ${entityType} ${entityId}, skipping`);
+            return null;
+        }
+
+        try {
+            const content = entity.getContent();
+            if (!content || !(content instanceof Buffer)) {
+                throw new Error(`Cannot get content for ${entityType} ${entityId}`);
+            }
+
+            const language = this.resolveOcrLanguage(languageNoteId, options.language);
+            const ocrResult = await this.extractTextFromFile(content, mime, { ...options, language });
+
+            this.storeOCRResult(blobId, ocrResult);
+
+            return ocrResult;
+        } catch (error) {
+            log.error(`Failed to process OCR for ${entityType} ${entityId}: ${error}`);
+            throw error;
+        }
+    }
+
+    /**
+     * Store OCR result in blob
+     */
+    storeOCRResult(blobId: string | undefined, ocrResult: OCRResult): void {
+        if (!blobId) {
+            log.error('Cannot store OCR result: blobId is undefined');
+            return;
+        }
+
+        try {
+            sql.execute(`
+                UPDATE blobs SET textRepresentation = ?
+                WHERE blobId = ?
+            `, [ocrResult.text, blobId]);
+
+            this.putBlobEntityChange(blobId);
+
+            log.info(`Stored OCR result for blob ${blobId}`);
+        } catch (error) {
+            log.error(`Failed to store OCR result for blob ${blobId}: ${error}`);
+            throw error;
+        }
+    }
+
+    /**
+     * Check whether a blob already has a stored text representation.
+     */
+    private hasStoredOCRResult(blobId: string | undefined): boolean {
+        if (!blobId) {
+            return false;
+        }
+
+        const row = sql.getRow<{ textRepresentation: string | null }>(
+            `SELECT textRepresentation FROM blobs WHERE blobId = ?`,
+            [blobId]
+        );
+
+        return !!row?.textRepresentation;
+    }
+
+    // Batch processing state
+    private batchProcessingState: {
+        inProgress: boolean;
+        total: number;
+        processed: number;
+        startTime?: Date;
+    } = {
+        inProgress: false,
+        total: 0,
+        processed: 0
+    };
+
+    /**
+     * Start batch OCR processing with progress tracking
+     */
+    async startBatchProcessing(): Promise<{ success: boolean; message?: string }> {
+        if (this.batchProcessingState.inProgress) {
+            return { success: false, message: 'Batch processing already in progress' };
+        }
+
+        try {
+            const blobsNeedingOCR = this.getBlobsNeedingOCR();
+
+            if (blobsNeedingOCR.length === 0) {
+                return { success: false, message: 'No images found that need OCR processing' };
+            }
+
+            this.batchProcessingState = {
+                inProgress: true,
+                total: blobsNeedingOCR.length,
+                processed: 0,
+                startTime: new Date()
+            };
+
+            // Start processing in background
+            this.processBlobs(blobsNeedingOCR).catch(error => {
+                log.error(`Batch processing failed: ${error instanceof Error ? error.message : String(error)}`);
+            }).finally(() => {
+                this.batchProcessingState.inProgress = false;
+            });
+
+            return { success: true };
+        } catch (error) {
+            log.error(`Failed to start batch processing: ${error instanceof Error ? error.message : String(error)}`);
+            return { success: false, message: error instanceof Error ? error.message : String(error) };
+        }
+    }
+
+    /**
+     * Get batch processing progress
+     */
+    getBatchProgress(): { inProgress: boolean; total: number; processed: number; percentage?: number; startTime?: Date } {
+        const result: { inProgress: boolean; total: number; processed: number; percentage?: number; startTime?: Date } = { ...this.batchProcessingState };
+        if (result.total > 0) {
+            result.percentage = (result.processed / result.total) * 100;
+        }
+        return result;
+    }
+
+    /**
+     * Cancel batch processing
+     */
+    cancelBatchProcessing(): void {
+        if (this.batchProcessingState.inProgress) {
+            this.batchProcessingState.inProgress = false;
+            log.info('Batch OCR processing cancelled');
+        }
+    }
+
+    /**
+     * Process a list of blobs sequentially, updating batch progress.
+     */
+    private async processBlobs(blobs: Array<{ entityType: 'note' | 'attachment'; entityId: string }>): Promise<void> {
+        log.info(`Starting batch OCR processing of ${blobs.length} items...`);
+
+        for (const blob of blobs) {
+            if (!this.batchProcessingState.inProgress) {
+                break;
+            }
+
+            try {
+                await this.processOcrEntity(blob);
+            } catch (error) {
+                log.error(`Failed to process OCR for ${blob.entityType} ${blob.entityId}: ${error}`);
+            }
+
+            this.batchProcessingState.processed++;
+
+            // Small delay to prevent overwhelming the system
+            await new Promise(resolve => setTimeout(resolve, 500));
+        }
+
+        log.info(`Batch OCR processing completed. Processed ${this.batchProcessingState.processed} files.`);
+    }
+
+    /**
+     * Process OCR for a single entity (note or attachment) by type.
+     */
+    private async processOcrEntity(entity: { entityType: 'note' | 'attachment'; entityId: string }): Promise<void> {
+        if (entity.entityType === 'note') {
+            await this.processNoteOCR(entity.entityId);
+        } else {
+            await this.processAttachmentOCR(entity.entityId);
+        }
+    }
+
+    /**
+     * Get processor for a given MIME type
+     */
+    /**
+     * Notifies the sync system that a blob has changed, without modifying the blob's identity.
+     */
+    private putBlobEntityChange(blobId: string): void {
+        const blob = becca.getBlob({ blobId });
+        if (!blob || !blob.blobId) return;
+
+        const hash = blobService.calculateContentHash({
+            blobId: blob.blobId,
+            content: blob.content,
+            textRepresentation: blob.textRepresentation,
+            utcDateModified: blob.utcDateModified!
+        });
+        entityChangesService.putEntityChange({
+            entityName: "blobs",
+            entityId: blobId,
+            hash,
+            isErased: false,
+            utcDateChanged: blob.utcDateModified,
+            isSynced: true
+        });
+    }
+
+    private getProcessorForMimeType(mimeType: string): FileProcessor | null {
+        for (const processor of this.processors.values()) {
+            if (processor.canProcess(mimeType)) {
+                return processor;
+            }
+        }
+        return null;
+    }
+
+    /**
+     * Get all MIME types supported by all registered processors
+     */
+    getAllSupportedMimeTypes(): string[] {
+        const supportedTypes = new Set<string>();
+
+        // Gather MIME types from all registered processors
+        for (const processor of this.processors.values()) {
+            const processorTypes = processor.getSupportedMimeTypes();
+            processorTypes.forEach(type => supportedTypes.add(type));
+        }
+
+        return Array.from(supportedTypes);
+    }
+
+
+    /**
+     * Get blobs that need OCR processing (those without text representation)
+     */
+    getBlobsNeedingOCR(): Array<{ blobId: string; mimeType: string; entityType: 'note' | 'attachment'; entityId: string }> {
+        try {
+            const supportedMimes = this.getAllSupportedMimeTypes();
+            const placeholders = supportedMimes.map(() => '?').join(', ');
+
+            const noteBlobs = sql.getRows<{
+                blobId: string;
+                mimeType: string;
+                entityId: string;
+            }>(`
+                SELECT n.blobId, n.mime as mimeType, n.noteId as entityId
+                FROM notes n
+                JOIN blobs b ON n.blobId = b.blobId
+                WHERE (n.type = 'image' OR (n.type = 'file' AND n.mime IN (${placeholders})))
+                AND n.isDeleted = 0
+                AND n.blobId IS NOT NULL
+                AND b.textRepresentation IS NULL
+            `, supportedMimes);
+
+            const attachmentBlobs = sql.getRows<{
+                blobId: string;
+                mimeType: string;
+                entityId: string;
+            }>(`
+                SELECT a.blobId, a.mime as mimeType, a.attachmentId as entityId
+                FROM attachments a
+                JOIN blobs b ON a.blobId = b.blobId
+                WHERE (a.role = 'image' OR (a.role = 'file' AND a.mime IN (${placeholders})))
+                AND a.isDeleted = 0
+                AND a.blobId IS NOT NULL
+                AND b.textRepresentation IS NULL
+            `, supportedMimes);
+
+            // Combine results
+            const result = [
+                ...noteBlobs.map(blob => ({ ...blob, entityType: 'note' as const })),
+                ...attachmentBlobs.map(blob => ({ ...blob, entityType: 'attachment' as const }))
+            ];
+
+            // Return all results (no need to filter by MIME type as we already did in the query)
+            return result;
+        } catch (error) {
+            log.error(`Failed to get blobs needing OCR: ${error}`);
+            return [];
+        }
+    }
+
+}
+
+export default new OCRService();
--- a/apps/server/src/services/ocr/processors/file_processor.ts
+++ b/apps/server/src/services/ocr/processors/file_processor.ts
@@ -0,0 +1,26 @@
+import { OCRResult, OCRProcessingOptions } from '../ocr_service.js';
+
+/**
+ * Base class for file processors that extract text from different file types
+ */
+export abstract class FileProcessor {
+    /**
+     * Check if this processor can handle the given MIME type
+     */
+    abstract canProcess(mimeType: string): boolean;
+
+    /**
+     * Extract text from the given file buffer
+     */
+    abstract extractText(buffer: Buffer, options: OCRProcessingOptions): Promise<OCRResult>;
+
+    /**
+     * Get the processing type identifier
+     */
+    abstract getProcessingType(): string;
+
+    /**
+     * Get list of MIME types supported by this processor
+     */
+    abstract getSupportedMimeTypes(): string[];
+}
--- a/apps/server/src/services/ocr/processors/image_processor.ts
+++ b/apps/server/src/services/ocr/processors/image_processor.ts
@@ -0,0 +1,160 @@
+import fs from 'fs';
+import Tesseract from 'tesseract.js';
+
+import dataDirs from '../../data_dir.js';
+import log from '../../log.js';
+import options from '../../options.js';
+import { OCRProcessingOptions,OCRResult } from '../ocr_service.js';
+import { FileProcessor } from './file_processor.js';
+
+/**
+ * Image processor for extracting text from image files using Tesseract
+ */
+export class ImageProcessor extends FileProcessor {
+    private worker: Tesseract.Worker | null = null;
+    private currentLanguage: string | null = null;
+    private readonly supportedTypes = [
+        'image/jpeg',
+        'image/jpg',
+        'image/png',
+        'image/gif',
+        'image/bmp',
+        'image/tiff',
+        'image/webp'
+    ];
+
+    canProcess(mimeType: string): boolean {
+        return this.supportedTypes.includes(mimeType.toLowerCase());
+    }
+
+    getSupportedMimeTypes(): string[] {
+        return [...this.supportedTypes];
+    }
+
+    async extractText(buffer: Buffer, options: OCRProcessingOptions = {}): Promise<OCRResult> {
+        const language = options.language || "eng";
+        await this.ensureWorker(language);
+
+        try {
+            log.info(`Starting image OCR text extraction (language: ${language})...`);
+
+            const result = await this.worker!.recognize(buffer);
+
+            // Filter text based on minimum confidence threshold
+            const { filteredText, overallConfidence } = this.filterTextByConfidence(result.data);
+
+            const ocrResult: OCRResult = {
+                text: filteredText,
+                confidence: overallConfidence,
+                extractedAt: new Date().toISOString(),
+                language,
+                pageCount: 1
+            };
+
+            return ocrResult;
+
+        } catch (error) {
+            log.error(`Image OCR text extraction failed: ${error}`);
+            throw error;
+        }
+    }
+
+    getProcessingType(): string {
+        return 'image';
+    }
+
+    /**
+     * Ensures a Tesseract worker is ready for the given language.
+     * Creates a new worker if none exists or if the language has changed.
+     */
+    private async ensureWorker(language: string): Promise<void> {
+        if (this.worker && this.currentLanguage === language) {
+            return;
+        }
+
+        if (this.worker) {
+            await this.worker.terminate();
+        }
+
+        fs.mkdirSync(dataDirs.OCR_CACHE_DIR, { recursive: true });
+
+        log.info(`Initializing Tesseract worker for language(s): ${language}`);
+        this.worker = await Tesseract.createWorker(language, 1, {
+            cachePath: dataDirs.OCR_CACHE_DIR,
+            logger: (m: { status: string; progress: number }) => {
+                if (m.status === 'recognizing text') {
+                    log.info(`Image OCR progress (${language}): ${Math.round(m.progress * 100)}%`);
+                }
+            }
+        });
+        this.currentLanguage = language;
+    }
+
+
+    /**
+     * Filter text based on minimum confidence threshold
+     */
+    private filterTextByConfidence(data: any): { filteredText: string; overallConfidence: number } {
+        const minConfidence = this.getMinConfidenceThreshold();
+
+        // If no minimum confidence set, return original text
+        if (minConfidence <= 0) {
+            return {
+                filteredText: data.text.trim(),
+                overallConfidence: data.confidence / 100
+            };
+        }
+
+        const filteredWords: string[] = [];
+        const validConfidences: number[] = [];
+
+        // Tesseract provides word-level data
+        if (data.words && Array.isArray(data.words)) {
+            for (const word of data.words) {
+                const wordConfidence = word.confidence / 100; // Convert to decimal
+
+                if (wordConfidence >= minConfidence) {
+                    filteredWords.push(word.text);
+                    validConfidences.push(wordConfidence);
+                }
+            }
+        } else {
+            // Fallback: if word-level data not available, use overall confidence
+            const overallConfidence = data.confidence / 100;
+            if (overallConfidence >= minConfidence) {
+                return {
+                    filteredText: data.text.trim(),
+                    overallConfidence
+                };
+            }
+            log.info(`Entire text filtered out due to low confidence ${overallConfidence} (below threshold ${minConfidence})`);
+            return {
+                filteredText: '',
+                overallConfidence
+            };
+        }
+
+        // Calculate average confidence of accepted words
+        const averageConfidence = validConfidences.length > 0
+            ? validConfidences.reduce((sum, conf) => sum + conf, 0) / validConfidences.length
+            : 0;
+
+        const filteredText = filteredWords.join(' ').trim();
+
+        log.info(`Filtered OCR text: ${filteredWords.length} words kept out of ${data.words?.length || 0} total words (min confidence: ${minConfidence})`);
+
+        return {
+            filteredText,
+            overallConfidence: averageConfidence
+        };
+    }
+
+    /**
+     * Get minimum confidence threshold from options
+     */
+    private getMinConfidenceThreshold(): number {
+        const minConfidence = options.getOption('ocrMinConfidence') ?? 0;
+        return parseFloat(minConfidence);
+    }
+
+}
--- a/apps/server/src/services/ocr/processors/office_processor.ts
+++ b/apps/server/src/services/ocr/processors/office_processor.ts
@@ -0,0 +1,70 @@
+import { parseExcel } from 'officeparser/dist/parsers/ExcelParser.js';
+import { parseOpenOffice } from 'officeparser/dist/parsers/OpenOfficeParser.js';
+import { parsePowerPoint } from 'officeparser/dist/parsers/PowerPointParser.js';
+import { parseWord } from 'officeparser/dist/parsers/WordParser.js';
+import type { OfficeParserConfig } from 'officeparser/dist/types.js';
+
+import log from '../../log.js';
+import { OCRProcessingOptions, OCRResult } from '../ocr_service.js';
+import { FileProcessor } from './file_processor.js';
+
+type Parser = (buffer: Buffer, config: OfficeParserConfig) => Promise<{ toText(): string }>;
+
+const PARSER_BY_MIME: Record<string, Parser> = {
+    // Office Open XML
+    'application/vnd.openxmlformats-officedocument.wordprocessingml.document': parseWord,
+    'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet': parseExcel,
+    'application/vnd.openxmlformats-officedocument.presentationml.presentation': parsePowerPoint,
+    // OpenDocument
+    'application/vnd.oasis.opendocument.text': parseOpenOffice,
+    'application/vnd.oasis.opendocument.spreadsheet': parseOpenOffice,
+    'application/vnd.oasis.opendocument.presentation': parseOpenOffice
+};
+
+const PARSER_CONFIG: OfficeParserConfig = {
+    outputErrorToConsole: false,
+    newlineDelimiter: '\n',
+    ignoreNotes: false,
+    putNotesAtLast: false
+};
+
+/**
+ * Office document processor for extracting text from DOCX/XLSX/PPTX and ODT/ODS/ODP files.
+ * Uses individual parsers from officeparser v6 to avoid pulling in pdfjs-dist.
+ */
+export class OfficeProcessor extends FileProcessor {
+
+    canProcess(mimeType: string): boolean {
+        return mimeType in PARSER_BY_MIME;
+    }
+
+    getSupportedMimeTypes(): string[] {
+        return Object.keys(PARSER_BY_MIME);
+    }
+
+    async extractText(buffer: Buffer, options: OCRProcessingOptions = {}): Promise<OCRResult> {
+        const mimeType = options.mimeType;
+        if (!mimeType || !(mimeType in PARSER_BY_MIME)) {
+            throw new Error(`Unsupported MIME type for Office processor: ${mimeType}`);
+        }
+
+        log.info(`Starting Office document text extraction for ${mimeType}...`);
+
+        const parse = PARSER_BY_MIME[mimeType];
+        const ast = await parse(buffer, PARSER_CONFIG);
+        const trimmed = ast.toText().trim();
+
+        return {
+            text: trimmed,
+            confidence: trimmed.length > 0 ? 0.99 : 0,
+            extractedAt: new Date().toISOString(),
+            language: options.language || "eng",
+            pageCount: 1
+        };
+    }
+
+    getProcessingType(): string {
+        return 'office';
+    }
+
+}
--- a/apps/server/src/services/ocr/processors/pdf_processor.ts
+++ b/apps/server/src/services/ocr/processors/pdf_processor.ts
@@ -0,0 +1,39 @@
+import { extractText, getDocumentProxy } from 'unpdf';
+
+import log from '../../log.js';
+import { OCRProcessingOptions, OCRResult } from '../ocr_service.js';
+import { FileProcessor } from './file_processor.js';
+
+/**
+ * PDF processor for extracting embedded text from PDF files using unpdf.
+ */
+export class PDFProcessor extends FileProcessor {
+
+    canProcess(mimeType: string): boolean {
+        return mimeType.toLowerCase() === 'application/pdf';
+    }
+
+    getSupportedMimeTypes(): string[] {
+        return ['application/pdf'];
+    }
+
+    async extractText(buffer: Buffer, options: OCRProcessingOptions = {}): Promise<OCRResult> {
+        log.info('Starting PDF text extraction...');
+
+        const pdf = await getDocumentProxy(new Uint8Array(buffer));
+        const { totalPages, text } = await extractText(pdf, { mergePages: true });
+
+        return {
+            text: text.trim(),
+            confidence: 0.99,
+            extractedAt: new Date().toISOString(),
+            language: options.language || "eng",
+            pageCount: totalPages
+        };
+    }
+
+    getProcessingType(): string {
+        return 'pdf';
+    }
+
+}
--- a/apps/server/src/services/options_init.ts
+++ b/apps/server/src/services/options_init.ts
@@ -212,7 +212,11 @@ const defaultOptions: DefaultOption[] = [
    { name: "experimentalFeatures", value: "[]", isSynced: true },

    // AI / LLM
-    { name: "llmProviders", value: "[]", isSynced: false }
+    { name: "llmProviders", value: "[]", isSynced: false },
+
+    // OCR options
+    { name: "ocrAutoProcessImages", value: "false", isSynced: true },
+    { name: "ocrMinConfidence", value: "0.75", isSynced: true },
 ];

 /**
--- a/apps/server/src/services/search/expressions/ocr_content.ts
+++ b/apps/server/src/services/search/expressions/ocr_content.ts
@@ -0,0 +1,80 @@
+import becca from "../../../becca/becca.js";
+import sql from "../../sql.js";
+import NoteSet from "../note_set.js";
+import type SearchContext from "../search_context.js";
+import Expression from "./expression.js";
+
+/**
+ * Search expression for finding text within OCR-extracted content (textRepresentation)
+ * from image notes and their attachments.
+ *
+ * Uses a single SQL query to find all noteIds whose own blob or attachment blobs
+ * contain matching text, then intersects with the input note set.
+ */
+export default class OCRContentExpression extends Expression {
+    private tokens: string[];
+
+    constructor(tokens: string[]) {
+        super();
+        this.tokens = tokens;
+    }
+
+    execute(inputNoteSet: NoteSet, executionContext: object, searchContext: SearchContext): NoteSet {
+        const resultNoteSet = new NoteSet();
+        const matchingNoteIds = this.findNoteIdsWithMatchingOCR();
+
+        for (const noteId of matchingNoteIds) {
+            const note = becca.notes[noteId];
+            if (note && inputNoteSet.hasNoteId(noteId)) {
+                resultNoteSet.add(note);
+            }
+        }
+
+        if (resultNoteSet.notes.length > 0) {
+            const highlightTokens = this.tokens
+                .filter(token => token.length > 2)
+                .map(token => token.toLowerCase());
+            searchContext.highlightedTokens.push(...highlightTokens);
+        }
+
+        return resultNoteSet;
+    }
+
+    /**
+     * Find all noteIds that have OCR text matching all tokens, in a single query.
+     * Checks both the note's own blob and its attachment blobs.
+     */
+    private findNoteIdsWithMatchingOCR(): Set<string> {
+        if (this.tokens.length === 0) return new Set();
+
+        // Build WHERE conditions: all tokens must appear in textRepresentation
+        const likeConditions = this.tokens.map(() => `b.textRepresentation LIKE ?`).join(' AND ');
+        const params = this.tokens.map(token => `%${token}%`);
+
+        // Find notes whose own blob matches
+        const noteIds = sql.getColumn<string>(`
+            SELECT n.noteId
+            FROM notes n
+            JOIN blobs b ON n.blobId = b.blobId
+            WHERE b.textRepresentation IS NOT NULL
+              AND n.isDeleted = 0
+              AND ${likeConditions}
+        `, params);
+
+        // Find notes that own attachments whose blob matches
+        const attachmentOwnerIds = sql.getColumn<string>(`
+            SELECT a.ownerId
+            FROM attachments a
+            JOIN blobs b ON a.blobId = b.blobId
+            WHERE b.textRepresentation IS NOT NULL
+              AND a.isDeleted = 0
+              AND ${likeConditions}
+        `, params);
+
+        return new Set([...noteIds, ...attachmentOwnerIds]);
+    }
+
+    toString(): string {
+        return `OCRContent('${this.tokens.join("', '")}')`;
+    }
+}
--- a/apps/server/src/services/search/search_result.ts
+++ b/apps/server/src/services/search/search_result.ts
@@ -1,12 +1,9 @@
-"use strict";
-
-import beccaService from "../../becca/becca_service.js";
 import becca from "../../becca/becca.js";
-import { 
-    normalizeSearchText, 
-    calculateOptimizedEditDistance, 
-    FUZZY_SEARCH_CONFIG 
-} from "./utils/text_utils.js";
+import beccaService from "../../becca/becca_service.js";
+import {
+    calculateOptimizedEditDistance,
+    FUZZY_SEARCH_CONFIG,
+    normalizeSearchText} from "./utils/text_utils.js";

 // Scoring constants for better maintainability
 const SCORE_WEIGHTS = {
@@ -98,7 +95,7 @@ class SearchResult {
        for (const chunk of chunks) {
            for (const token of tokens) {
                const normalizedToken = normalizeSearchText(token.toLowerCase());
-                
+
                if (chunk === normalizedToken) {
                    tokenScore += SCORE_WEIGHTS.TOKEN_EXACT_MATCH * token.length * factor;
                } else if (chunk.startsWith(normalizedToken)) {
@@ -108,10 +105,10 @@ class SearchResult {
                } else {
                    // Try fuzzy matching for individual tokens with caps applied
                    const editDistance = calculateOptimizedEditDistance(chunk, normalizedToken, FUZZY_SEARCH_CONFIG.MAX_EDIT_DISTANCE);
-                    if (editDistance <= FUZZY_SEARCH_CONFIG.MAX_EDIT_DISTANCE && 
+                    if (editDistance <= FUZZY_SEARCH_CONFIG.MAX_EDIT_DISTANCE &&
                        normalizedToken.length >= FUZZY_SEARCH_CONFIG.MIN_FUZZY_TOKEN_LENGTH &&
                        this.fuzzyScore < SCORE_WEIGHTS.MAX_TOTAL_FUZZY_SCORE) {
-                        
+
                        const fuzzyWeight = SCORE_WEIGHTS.TOKEN_FUZZY_MATCH * (1 - editDistance / FUZZY_SEARCH_CONFIG.MAX_EDIT_DISTANCE);
                        // Apply caps: limit token length multiplier and per-token contribution
                        const cappedTokenLength = Math.min(token.length, SCORE_WEIGHTS.MAX_FUZZY_TOKEN_LENGTH_MULTIPLIER);
@@ -119,7 +116,7 @@ class SearchResult {
                            fuzzyWeight * cappedTokenLength * factor,
                            SCORE_WEIGHTS.MAX_FUZZY_SCORE_PER_TOKEN
                        );
-                        
+
                        tokenScore += fuzzyTokenScore;
                        this.fuzzyScore += fuzzyTokenScore;
                    }
@@ -129,13 +126,12 @@ class SearchResult {
        this.score += tokenScore;
    }

-
    /**
     * Checks if the query matches as a complete word in the text
     */
    private isWordMatch(text: string, query: string): boolean {
-        return text.includes(` ${query} `) || 
-               text.startsWith(`${query} `) || 
+        return text.includes(` ${query} `) ||
+               text.startsWith(`${query} `) ||
               text.endsWith(` ${query}`);
    }

@@ -147,21 +143,21 @@ class SearchResult {
        if (this.fuzzyScore >= SCORE_WEIGHTS.MAX_TOTAL_FUZZY_SCORE) {
            return 0;
        }
-        
+
        const editDistance = calculateOptimizedEditDistance(title, query, FUZZY_SEARCH_CONFIG.MAX_EDIT_DISTANCE);
        const maxLen = Math.max(title.length, query.length);
-        
+
        // Only apply fuzzy matching if the query is reasonably long and edit distance is small
-        if (query.length >= FUZZY_SEARCH_CONFIG.MIN_FUZZY_TOKEN_LENGTH && 
-            editDistance <= FUZZY_SEARCH_CONFIG.MAX_EDIT_DISTANCE && 
+        if (query.length >= FUZZY_SEARCH_CONFIG.MIN_FUZZY_TOKEN_LENGTH &&
+            editDistance <= FUZZY_SEARCH_CONFIG.MAX_EDIT_DISTANCE &&
            editDistance / maxLen <= 0.3) {
            const similarity = 1 - (editDistance / maxLen);
            const baseFuzzyScore = SCORE_WEIGHTS.TITLE_WORD_MATCH * similarity * 0.7; // Reduced weight for fuzzy matches
-            
+
            // Apply cap to ensure fuzzy title matches don't exceed reasonable bounds
            return Math.min(baseFuzzyScore, SCORE_WEIGHTS.MAX_TOTAL_FUZZY_SCORE * 0.3);
        }
-        
+
        return 0;
    }

--- a/apps/server/src/services/search/search_result_ocr.spec.ts
+++ b/apps/server/src/services/search/search_result_ocr.spec.ts
@@ -0,0 +1,149 @@
+import { beforeEach, describe, expect, it, vi } from 'vitest';
+
+const mockBecca = {
+    notes: {} as Record<string, any>,
+    getNote: vi.fn()
+};
+
+const mockBeccaService = {
+    getNoteTitleForPath: vi.fn()
+};
+
+vi.mock('../../becca/becca.js', () => ({
+    default: mockBecca
+}));
+
+vi.mock('../../becca/becca_service.js', () => ({
+    default: mockBeccaService
+}));
+
+let SearchResult: any;
+
+beforeEach(async () => {
+    vi.clearAllMocks();
+
+    mockBeccaService.getNoteTitleForPath.mockReturnValue('Test Note Title');
+
+    mockBecca.notes['test123'] = {
+        noteId: 'test123',
+        title: 'Test Note',
+        isInHiddenSubtree: vi.fn().mockReturnValue(false)
+    };
+
+    const module = await import('./search_result.js');
+    SearchResult = module.default;
+});
+
+describe('SearchResult', () => {
+    describe('constructor', () => {
+        it('should initialize with note path array', () => {
+            const searchResult = new SearchResult(['root', 'folder', 'test123']);
+
+            expect(searchResult.notePathArray).toEqual(['root', 'folder', 'test123']);
+            expect(searchResult.noteId).toBe('test123');
+            expect(searchResult.notePath).toBe('root/folder/test123');
+            expect(searchResult.score).toBe(0);
+            expect(mockBeccaService.getNoteTitleForPath).toHaveBeenCalledWith(['root', 'folder', 'test123']);
+        });
+    });
+
+    describe('computeScore', () => {
+        let searchResult: any;
+
+        beforeEach(() => {
+            searchResult = new SearchResult(['root', 'test123']);
+        });
+
+        describe('basic scoring', () => {
+            it('should give highest score for exact note ID match', () => {
+                searchResult.computeScore('test123', ['test123']);
+                expect(searchResult.score).toBeGreaterThanOrEqual(1000);
+            });
+
+            it('should give high score for exact title match', () => {
+                searchResult.computeScore('test note', ['test', 'note']);
+                expect(searchResult.score).toBeGreaterThan(2000);
+            });
+
+            it('should give medium score for title prefix match', () => {
+                searchResult.computeScore('test', ['test']);
+                expect(searchResult.score).toBeGreaterThan(500);
+            });
+
+            it('should give lower score for title word match', () => {
+                mockBecca.notes['test123'].title = 'This is a test note';
+                searchResult.computeScore('test', ['test']);
+                expect(searchResult.score).toBeGreaterThan(300);
+            });
+        });
+
+        describe('hidden notes penalty', () => {
+            it('should apply penalty for hidden notes', () => {
+                mockBecca.notes['test123'].isInHiddenSubtree.mockReturnValue(true);
+
+                searchResult.computeScore('test', ['test']);
+                const hiddenScore = searchResult.score;
+
+                mockBecca.notes['test123'].isInHiddenSubtree.mockReturnValue(false);
+                searchResult.score = 0;
+                searchResult.computeScore('test', ['test']);
+                const normalScore = searchResult.score;
+
+                expect(normalScore).toBeGreaterThan(hiddenScore);
+                expect(hiddenScore).toBe(normalScore / 3);
+            });
+        });
+    });
+
+    describe('addScoreForStrings', () => {
+        let searchResult: any;
+
+        beforeEach(() => {
+            searchResult = new SearchResult(['root', 'test123']);
+        });
+
+        it('should give highest score for exact token match', () => {
+            searchResult.addScoreForStrings(['sample'], 'sample text', 1.0);
+            const exactScore = searchResult.score;
+
+            searchResult.score = 0;
+            searchResult.addScoreForStrings(['sample'], 'sampling text', 1.0);
+            const prefixScore = searchResult.score;
+
+            searchResult.score = 0;
+            searchResult.addScoreForStrings(['sample'], 'text sample text', 1.0);
+            const partialScore = searchResult.score;
+
+            expect(exactScore).toBeGreaterThan(prefixScore);
+            expect(exactScore).toBeGreaterThanOrEqual(partialScore);
+        });
+
+        it('should apply factor multiplier correctly', () => {
+            searchResult.addScoreForStrings(['sample'], 'sample text', 2.0);
+            const doubleFactorScore = searchResult.score;
+
+            searchResult.score = 0;
+            searchResult.addScoreForStrings(['sample'], 'sample text', 1.0);
+            const singleFactorScore = searchResult.score;
+
+            expect(doubleFactorScore).toBe(singleFactorScore * 2);
+        });
+
+        it('should handle multiple tokens', () => {
+            searchResult.addScoreForStrings(['hello', 'world'], 'hello world test', 1.0);
+            expect(searchResult.score).toBeGreaterThan(0);
+        });
+
+        it('should be case insensitive', () => {
+            searchResult.addScoreForStrings(['sample'], 'sample text', 1.0);
+            const lowerCaseScore = searchResult.score;
+
+            searchResult.score = 0;
+            searchResult.addScoreForStrings(['sample'], 'SAMPLE text', 1.0);
+            const upperCaseScore = searchResult.score;
+
+            expect(upperCaseScore).toEqual(lowerCaseScore);
+            expect(upperCaseScore).toBeGreaterThan(0);
+        });
+    });
+});
--- a/apps/server/src/services/search/services/parse.ts
+++ b/apps/server/src/services/search/services/parse.ts
@@ -1,28 +1,30 @@
-"use strict";
+

 import { dayjs } from "@triliumnext/commons";
+
+import { removeDiacritic } from "../../utils.js";
+import AncestorExp from "../expressions/ancestor.js";
 import AndExp from "../expressions/and.js";
-import OrExp from "../expressions/or.js";
-import NotExp from "../expressions/not.js";
+import AttributeExistsExp from "../expressions/attribute_exists.js";
 import ChildOfExp from "../expressions/child_of.js";
 import DescendantOfExp from "../expressions/descendant_of.js";
-import ParentOfExp from "../expressions/parent_of.js";
-import RelationWhereExp from "../expressions/relation_where.js";
-import PropertyComparisonExp from "../expressions/property_comparison.js";
-import AttributeExistsExp from "../expressions/attribute_exists.js";
-import LabelComparisonExp from "../expressions/label_comparison.js";
-import NoteFlatTextExp from "../expressions/note_flat_text.js";
-import NoteContentFulltextExp from "../expressions/note_content_fulltext.js";
-import OrderByAndLimitExp from "../expressions/order_by_and_limit.js";
-import AncestorExp from "../expressions/ancestor.js";
-import buildComparator from "./build_comparator.js";
-import ValueExtractor from "../value_extractor.js";
-import { removeDiacritic } from "../../utils.js";
-import TrueExp from "../expressions/true.js";
-import IsHiddenExp from "../expressions/is_hidden.js";
-import type SearchContext from "../search_context.js";
-import type { TokenData, TokenStructure } from "./types.js";
 import type Expression from "../expressions/expression.js";
+import IsHiddenExp from "../expressions/is_hidden.js";
+import LabelComparisonExp from "../expressions/label_comparison.js";
+import NotExp from "../expressions/not.js";
+import NoteContentFulltextExp from "../expressions/note_content_fulltext.js";
+import NoteFlatTextExp from "../expressions/note_flat_text.js";
+import OCRContentExpression from "../expressions/ocr_content.js";
+import OrExp from "../expressions/or.js";
+import OrderByAndLimitExp from "../expressions/order_by_and_limit.js";
+import ParentOfExp from "../expressions/parent_of.js";
+import PropertyComparisonExp from "../expressions/property_comparison.js";
+import RelationWhereExp from "../expressions/relation_where.js";
+import TrueExp from "../expressions/true.js";
+import type SearchContext from "../search_context.js";
+import ValueExtractor from "../value_extractor.js";
+import buildComparator from "./build_comparator.js";
+import type { TokenData, TokenStructure } from "./types.js";

 function getFulltext(_tokens: TokenData[], searchContext: SearchContext, leadingOperator?: string) {
    const tokens: string[] = _tokens.map((t) => removeDiacritic(t.token));
@@ -42,16 +44,26 @@ function getFulltext(_tokens: TokenData[], searchContext: SearchContext, leading
            // Exact match on title OR exact match on content OR exact match in flat text (includes attributes)
            // For multi-word, join tokens with space to form exact phrase
            const titleSearchValue = tokens.join(" ");
-            return new OrExp([
+            const exactMatchExpressions: Expression[] = [
                new PropertyComparisonExp(searchContext, "title", "=", titleSearchValue),
                new NoteContentFulltextExp("=", { tokens, flatText: false }),
                new NoteContentFulltextExp("=", { tokens, flatText: true })
-            ]);
+            ];
+
+            exactMatchExpressions.push(new OCRContentExpression(tokens));
+
+            return new OrExp(exactMatchExpressions);
        }
-        return new OrExp([new NoteFlatTextExp(tokens), new NoteContentFulltextExp(operator, { tokens, flatText: true })]);
-    } else {
-        return new NoteFlatTextExp(tokens);
+
+        const searchExpressions: Expression[] = [
+            new NoteFlatTextExp(tokens),
+            new NoteContentFulltextExp(operator, { tokens, flatText: true }),
+            new OCRContentExpression(tokens)
+        ];
+
+        return new OrExp(searchExpressions);
    }
+    return new NoteFlatTextExp(tokens);
 }

 const OPERATORS = new Set(["=", "!=", "*=*", "*=", "=*", ">", ">=", "<", "<=", "%=", "~=", "~*"]);
@@ -298,9 +310,9 @@ function getExpression(tokens: TokenData[], searchContext: SearchContext, level
            searchContext.addError(`Relation can be compared only with property, e.g. ~relation.title=hello in ${context(i)}`);

            return null;
-        } else {
-            return new AttributeExistsExp("relation", relationName, searchContext.fuzzyAttributeSearch);
        }
+        return new AttributeExistsExp("relation", relationName, searchContext.fuzzyAttributeSearch);
+
    }

    function parseOrderByAndLimit() {
@@ -308,7 +320,7 @@ function getExpression(tokens: TokenData[], searchContext: SearchContext, level
            valueExtractor: ValueExtractor;
            direction: string;
        }[] = [];
-        let limit: number | undefined = undefined;
+        let limit: number | undefined;

        if (tokens[i].token === "orderby") {
            do {
@@ -354,9 +366,9 @@ function getExpression(tokens: TokenData[], searchContext: SearchContext, level
            return AndExp.of(expressions);
        } else if (op === "or") {
            return OrExp.of(expressions);
-        } else {
-            throw new Error(`Unrecognized op=${op}`);
        }
+        throw new Error(`Unrecognized op=${op}`);
+
    }

    for (i = 0; i < tokens.length; i++) {
@@ -423,7 +435,7 @@ function getExpression(tokens: TokenData[], searchContext: SearchContext, level
            } else if (op !== token) {
                searchContext.addError("Mixed usage of AND/OR - always use parenthesis to group AND/OR expressions.");
            }
-        } else if (isOperator({ token: token })) {
+        } else if (isOperator({ token })) {
            searchContext.addError(`Misplaced or incomplete expression "${token}"`);
        } else {
            searchContext.addError(`Unrecognized expression "${token}"`);
@@ -493,9 +505,9 @@ function getAncestorExp({ ancestorNoteId, ancestorDepth, includeHiddenNotes }: S
        return new AncestorExp(ancestorNoteId, ancestorDepth);
    } else if (!includeHiddenNotes) {
        return new NotExp(new IsHiddenExp());
-    } else {
-        return null;
    }
+    return null;
+
 }

 export default parse;
--- a/apps/server/src/services/search/services/search.ts
+++ b/apps/server/src/services/search/services/search.ts
@@ -435,21 +435,46 @@ function findFirstNoteWithQuery(query: string, searchContext: SearchContext): BN
    return searchResults.length > 0 ? becca.notes[searchResults[0].noteId] : null;
 }

+/**
+ * Returns the first non-empty textRepresentation for a note's own blob
+ * or any of its attachment blobs.
+ */
+function getTextRepresentationForNote(note: BNote): string | null {
+    // Query only textRepresentation to avoid loading large binary content into memory.
+    const row = sql.getRow<{ textRepresentation: string | null }>(`
+        SELECT b.textRepresentation FROM blobs b
+        WHERE b.textRepresentation IS NOT NULL
+          AND b.textRepresentation != ''
+          AND (
+              b.blobId = ?
+              OR b.blobId IN (SELECT blobId FROM attachments WHERE ownerId = ? AND isDeleted = 0)
+          )
+        LIMIT 1
+    `, [note.blobId, note.noteId]);
+
+    return row?.textRepresentation ?? null;
+}
+
 function extractContentSnippet(noteId: string, searchTokens: string[], maxLength: number = 200): string {
    const note = becca.notes[noteId];
    if (!note) {
        return "";
    }

-    // Only extract content for text-based notes
-    if (!["text", "code", "mermaid", "canvas", "mindMap"].includes(note.type)) {
-        return "";
-    }
-
    try {
-        let content = note.getContent();
-        
-        if (!content || typeof content !== "string") {
+        let content: string | undefined;
+
+        if (["text", "code", "mermaid", "canvas", "mindMap"].includes(note.type)) {
+            const raw = note.getContent();
+            if (raw && typeof raw === "string") {
+                content = raw;
+            }
+        } else {
+            // For non-text notes (image, file), use OCR text representation
+            content = getTextRepresentationForNote(note) || undefined;
+        }
+
+        if (!content) {
            return "";
        }

--- a/package.json
+++ b/package.json
@@ -36,7 +36,7 @@
    "test:all": "pnpm test:parallel && pnpm test:sequential",
    "test:parallel": "pnpm --filter=!server --filter=!ckeditor5-mermaid --filter=!ckeditor5-math --parallel test",
    "test:sequential": "pnpm --filter=server --filter=ckeditor5-mermaid --filter=ckeditor5-math --sequential test",
-    "typecheck": "tsc --build",
+    "typecheck": "tsx scripts/filter-tsc-output.mts",
    "dev:format-check": "eslint -c eslint.format.config.mjs .",
    "dev:format-fix": "eslint -c eslint.format.config.mjs . --fix",
    "dev:linter-check": "cross-env NODE_OPTIONS=--max_old_space_size=4096 eslint .",
--- a/packages/commons/src/lib/i18n.ts
+++ b/packages/commons/src/lib/i18n.ts
@@ -9,27 +9,29 @@ export interface Locale {
    devOnly?: boolean;
    /** The value to pass to `--lang` for the Electron instance in order to set it as a locale. Not setting it will hide it from the list of supported locales. */
    electronLocale?: "en" | "de" | "es" | "fr" | "zh_CN" | "zh_TW" | "ro" | "af" | "am" | "ar" | "bg" | "bn" | "ca" | "cs" | "da" | "el" | "en_GB" | "es_419" | "et" | "fa" | "fi" | "fil" | "gu" | "he" | "hi" | "hr" | "hu" | "id" | "it" | "ja" | "kn" | "ko" | "lt" | "lv" | "ml" | "mr" | "ms" | "nb" | "nl" | "pl" | "pt_BR" | "pt_PT" | "ru" | "sk" | "sl" | "sr" | "sv" | "sw" | "ta" | "te" | "th" | "tr" | "uk" | "ur" | "vi";
+    /** The Tesseract OCR language code for this locale (e.g. "eng", "fra", "deu"). See https://tesseract-ocr.github.io/tessdoc/Data-Files-in-different-versions.html */
+    tesseractCode?: "eng" | "deu" | "spa" | "fra" | "gle" | "ita" | "hin" | "jpn" | "por" | "pol" | "ron" | "rus" | "chi_sim" | "chi_tra" | "ukr" | "ara" | "heb" | "kur" | "fas" | "kor";
 }

 // When adding a new locale, prefer the version with hyphen instead of underscore.
 const UNSORTED_LOCALES = [
-    { id: "cn", name: "简体中文", electronLocale: "zh_CN" },
-    { id: "de", name: "Deutsch", electronLocale: "de" },
-    { id: "en", name: "English (United States)", electronLocale: "en" },
-    { id: "en-GB", name: "English (United Kingdom)", electronLocale: "en_GB" },
-    { id: "es", name: "Español", electronLocale: "es" },
-    { id: "fr", name: "Français", electronLocale: "fr" },
-    { id: "ga", name: "Gaeilge", electronLocale: "en" },
-    { id: "it", name: "Italiano", electronLocale: "it" },
-    { id: "hi", name: "हिन्दी", electronLocale: "hi" },
-    { id: "ja", name: "日本語", electronLocale: "ja" },
-    { id: "pt_br", name: "Português (Brasil)", electronLocale: "pt_BR" },
-    { id: "pt", name: "Português (Portugal)", electronLocale: "pt_PT" },
-    { id: "pl", name: "Polski", electronLocale: "pl" },
-    { id: "ro", name: "Română", electronLocale: "ro" },
-    { id: "ru", name: "Русский", electronLocale: "ru" },
-    { id: "tw", name: "繁體中文", electronLocale: "zh_TW" },
-    { id: "uk", name: "Українська", electronLocale: "uk" },
+    { id: "cn", name: "简体中文", electronLocale: "zh_CN", tesseractCode: "chi_sim" },
+    { id: "de", name: "Deutsch", electronLocale: "de", tesseractCode: "deu" },
+    { id: "en", name: "English (United States)", electronLocale: "en", tesseractCode: "eng" },
+    { id: "en-GB", name: "English (United Kingdom)", electronLocale: "en_GB", tesseractCode: "eng" },
+    { id: "es", name: "Español", electronLocale: "es", tesseractCode: "spa" },
+    { id: "fr", name: "Français", electronLocale: "fr", tesseractCode: "fra" },
+    { id: "ga", name: "Gaeilge", electronLocale: "en", tesseractCode: "gle" },
+    { id: "it", name: "Italiano", electronLocale: "it", tesseractCode: "ita" },
+    { id: "hi", name: "हिन्दी", electronLocale: "hi", tesseractCode: "hin" },
+    { id: "ja", name: "日本語", electronLocale: "ja", tesseractCode: "jpn" },
+    { id: "pt_br", name: "Português (Brasil)", electronLocale: "pt_BR", tesseractCode: "por" },
+    { id: "pt", name: "Português (Portugal)", electronLocale: "pt_PT", tesseractCode: "por" },
+    { id: "pl", name: "Polski", electronLocale: "pl", tesseractCode: "pol" },
+    { id: "ro", name: "Română", electronLocale: "ro", tesseractCode: "ron" },
+    { id: "ru", name: "Русский", electronLocale: "ru", tesseractCode: "rus" },
+    { id: "tw", name: "繁體中文", electronLocale: "zh_TW", tesseractCode: "chi_tra" },
+    { id: "uk", name: "Українська", electronLocale: "uk", tesseractCode: "ukr" },

    /**
     * Development-only languages.
@@ -53,25 +55,29 @@ const UNSORTED_LOCALES = [
        id: "ar",
        name: "اَلْعَرَبِيَّةُ",
        rtl: true,
-        electronLocale: "ar"
+        electronLocale: "ar",
+        tesseractCode: "ara"
    },
    { // Hebrew
        id: "he",
        name: "עברית",
        rtl: true,
-        contentOnly: true
+        contentOnly: true,
+        tesseractCode: "heb"
    },
    { // Kurdish
        id: "ku",
        name: "کوردی",
        rtl: true,
-        contentOnly: true
+        contentOnly: true,
+        tesseractCode: "kur"
    },
    { // Persian
        id: "fa",
        name: "فارسی",
        rtl: true,
-        contentOnly: true
+        contentOnly: true,
+        tesseractCode: "fas"
    }
 ] as const;

@@ -82,3 +88,10 @@ export const LOCALES: Locale[] = Array.from(UNSORTED_LOCALES)
 export type LOCALE_IDS = typeof UNSORTED_LOCALES[number]["id"];
 /** A type containing a string union of all the supported locales that are not content-only (i.e. can be used as the UI language). */
 export type DISPLAYABLE_LOCALE_IDS = Exclude<typeof UNSORTED_LOCALES[number], { contentOnly: true }>["id"];
+
+/**
+ * Returns the Tesseract OCR language code for the given locale ID, or `null` if not mapped.
+ */
+export function getTesseractCode(localeId: string): string | null {
+    return LOCALES.find((l) => l.id === localeId)?.tesseractCode ?? null;
+}
--- a/packages/commons/src/lib/options_interface.ts
+++ b/packages/commons/src/lib/options_interface.ts
@@ -144,6 +144,12 @@ export interface OptionDefinitions extends KeyboardShortcutsOptions<KeyboardActi
    // AI / LLM
    /** JSON array of configured LLM providers with their API keys */
    llmProviders: string;
+
+    // OCR options
+    ocrEnabled: boolean;
+    ocrLanguage: string;
+    ocrAutoProcessImages: boolean;
+    ocrMinConfidence: string;
 }

 export type OptionNames = keyof OptionDefinitions;
--- a/packages/commons/src/lib/rows.ts
+++ b/packages/commons/src/lib/rows.ts
@@ -72,6 +72,7 @@ export interface BlobRow {
    blobId: string;
    content: string | Buffer;
    contentLength: number;
+    textRepresentation?: string | null;
    dateModified: string;
    utcDateModified: string;
 }
--- a/packages/commons/src/lib/server_api.ts
+++ b/packages/commons/src/lib/server_api.ts
@@ -288,6 +288,13 @@ export interface ToMarkdownResponse {
    markdownContent: string;
 }

+export interface TextRepresentationResponse {
+    success: boolean;
+    text: string;
+    hasOcr: boolean;
+    message?: string;
+}
+
 export interface IconRegistry {
    sources: {
        prefix: string;
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -580,6 +580,9 @@ importers:
      sucrase:
        specifier: 3.35.1
        version: 3.35.1
+      unpdf:
+        specifier: 1.4.0
+        version: 1.4.0(@napi-rs/canvas@0.1.96)
    devDependencies:
      '@braintree/sanitize-url':
        specifier: 7.1.2
@@ -806,6 +809,9 @@ importers:
      normalize-strings:
        specifier: 1.1.1
        version: 1.1.1
+      officeparser:
+        specifier: 6.0.7
+        version: 6.0.7(encoding@0.1.13)
      rand-token:
        specifier: 1.0.1
        version: 1.0.1
@@ -839,6 +845,9 @@ importers:
      swagger-jsdoc:
        specifier: 6.2.8
        version: 6.2.8(openapi-types@12.1.3)
+      tesseract.js:
+        specifier: 6.0.1
+        version: 6.0.1(encoding@0.1.13)
      time2fa:
        specifier: 1.4.2
        version: 1.4.2
@@ -2572,9 +2581,6 @@ packages:
  '@emnapi/core@1.9.0':
    resolution: {integrity: sha512-0DQ98G9ZQZOxfUcQn1waV2yS8aWdZ6kJMbYCJB3oUBecjWYO1fqJ+a1DRfPF3O5JEkwqwP1A9QEN/9mYm2Yd0w==}

-  '@emnapi/runtime@1.8.1':
-    resolution: {integrity: sha512-mehfKSMWjjNol8659Z8KxEMrdSJDDot5SXMq00dM8BN4o+CLNXQ0xH2V7EchNHV4RmbZLmmPdEaXZc5H2FXmDg==}
-
  '@emnapi/runtime@1.9.0':
    resolution: {integrity: sha512-QN75eB0IH2ywSpRpNddCRfQIhmJYBCJ1x5Lb3IscKAL8bMnVAKnRg8dCoXbHzVLLH7P38N2Z3mtulB7W0J0FKw==}

@@ -7268,6 +7274,10 @@ packages:
    engines: {node: '>=10.0.0'}
    deprecated: this version has critical issues, please update to the latest version

+  '@xmldom/xmldom@0.8.12':
+    resolution: {integrity: sha512-9k/gHF6n/pAi/9tqr3m3aqkuiNosYTurLLUtc7xQ9sxB/wm7WPygCv8GYa6mS0fLJEHhqMC1ATYhz++U/lRHqg==}
+    engines: {node: '>=10.0.0'}
+
  '@xtuc/ieee754@1.2.0':
    resolution: {integrity: sha512-DX8nKgqcGwsc0eJSqYt5lwP4DH5FlHnmuWWBRy7X0NcaGR0ZtuyeESgMwTYVEtxmsNGY+qit4QYT/MIYTOTPeA==}

@@ -7766,6 +7776,9 @@ packages:
  blurhash@2.0.5:
    resolution: {integrity: sha512-cRygWd7kGBQO3VEhPiTgq4Wc43ctsM+o46urrmPOiuAe+07fzlSB9OJVdpgDL0jPqXUVQ9ht7aq7kxOeJHRK+w==}

+  bmp-js@0.1.0:
+    resolution: {integrity: sha512-vHdS19CnY3hwiNdkaqk93DvjVLfbEcI8mys4UjuWrlX1haDmroo8o4xCzh4wD6DGV6HxRCyauwhHRqMTfERtjw==}
+
  bmp-ts@1.0.9:
    resolution: {integrity: sha512-cTEHk2jLrPyi+12M3dhpEbnnPOsaZuq7C45ylbbQIiWgDFZq4UVYPEY5mlqjvsj/6gJv9qX5sa+ebDzLXT28Vw==}

@@ -10479,6 +10492,9 @@ packages:
    peerDependencies:
      postcss: ^8.1.0

+  idb-keyval@6.2.2:
+    resolution: {integrity: sha512-yjD9nARJ/jb1g+CvD0tlhUHOrJ9Sy0P8T9MF3YaLlHnSRpwPfpTX0XIvpmw3gAJUmEu3FiICLBDPXVwyEvrleg==}
+
  ieee754@1.2.1:
    resolution: {integrity: sha512-dcyqhDvX1C46lXZcVqCpK+FtMRQVdIMN6/Df5js2zouUsqG7I6sFxitIC+7KYK29KdXOLHdu9zL4sFnoVQnqaA==}

@@ -10876,6 +10892,9 @@ packages:
    resolution: {integrity: sha512-mE00Gnza5EEB3Ds0HfMyllZzbBrmLOX3vfWoj9A9PEnTfratQ/BcaJOuMhnkhjXvb2+FkY3VuHqtAGpTPmglFQ==}
    engines: {node: '>=18'}

+  is-url@1.2.4:
+    resolution: {integrity: sha512-ITvGim8FhRiYe4IQ5uHSkj7pVaPDrCTkNd3yq3cV7iZAcJdHTUMPMEHcqSOy9xZ9qFenQCvi+2wjH9a1nXqHww==}
+
  is-weakmap@2.0.2:
    resolution: {integrity: sha512-K5pXYOm9wqY1RgjpL3YTkF39tni1XajUIkawTLUo9EZEVUFga5gSQJF8nNS7ZwJQ02y+1YCNYcMh+HIf1ZqE+w==}
    engines: {node: '>= 0.4'}
@@ -12400,6 +12419,11 @@ packages:
  ofetch@1.5.1:
    resolution: {integrity: sha512-2W4oUZlVaqAPAil6FUg/difl6YhqhUR7x2eZY4bQCko22UXg3hptq9KLQdqFClV+Wu85UX7hNtdGTngi/1BxcA==}

+  officeparser@6.0.7:
+    resolution: {integrity: sha512-MkNHyWIfEZRDtB8c0fgJHdb4Ui0I/WztBjlUjlPiEbTO6dIYaJMt+llS5p5Foj13guUZgGxkkM9VwsVRthHNAA==}
+    engines: {node: '>=18.0.0'}
+    hasBin: true
+
  ohash@2.0.11:
    resolution: {integrity: sha512-RdR9FQrFwNBNXAr4GixM8YaRZRJ5PUWbKYbE5eOsrwAjJW0q2REGcf79oYPsLyskQCZG1PLN+S/K1V00joZAoQ==}

@@ -12458,6 +12482,10 @@ packages:
  openapi-types@12.1.3:
    resolution: {integrity: sha512-N4YtSYJqghVu4iek2ZUvcN/0aqH1kRDuNqzcycDxhOUpg7GdvLa2F3DgS6yBNhInhv2r/6I0Flkn7CqL8+nIcw==}

+  opencollective-postinstall@2.0.3:
+    resolution: {integrity: sha512-8AV/sCtuzUeTo8gQK5qDZzARrulB3egtLzFgteqB2tcT4Mw7B8Kt7JcDHmltjz6FOAHsvTevk70gZEbhM4ZS9Q==}
+    hasBin: true
+
  opener@1.5.2:
    resolution: {integrity: sha512-ur5UIdyw5Y7yEj9wLzhqXiy6GZ3Mwx0yGI+5sMn2r0N0v3cKJvUmFH5yPP+WXh9e0xfyzyJX95D8l088DNFj7A==}
    hasBin: true
@@ -13545,6 +13573,9 @@ packages:
  regenerate@1.4.2:
    resolution: {integrity: sha512-zrceR/XhGYU/d/opr2EKO7aRHUeiBI8qjtfHqADTwZd6Szfy16la6kqD0MIUs5z5hx6AaKa+PixpPrR289+I0A==}

+  regenerator-runtime@0.13.11:
+    resolution: {integrity: sha512-kY1AZVr2Ra+t+piVaJ4gxaFaReZVH40AKNo7UCX6W+dEwBo/2oZJzqfuN1qLq1oL45o56cPaTXELwrTh8Fpggg==}
+
  regexp-util@2.0.3:
    resolution: {integrity: sha512-GP6h9OgJmhAZpb3dbNbXTfRWVnGcoMhWRZv/HxgM4/qCVqs1P9ukQdYxaUhjWBSAs9oJ/uPXUUvGT1VMe0Bs0Q==}
    engines: {node: '>=16'}
@@ -14760,6 +14791,18 @@ packages:
    engines: {node: '>=10'}
    hasBin: true

+  tesseract.js-core@6.0.0:
+    resolution: {integrity: sha512-1Qncm/9oKM7xgrQXZXNB+NRh19qiXGhxlrR8EwFbK5SaUbPZnS5OMtP/ghtqfd23hsr1ZvZbZjeuAGcMxd/ooA==}
+
+  tesseract.js-core@7.0.0:
+    resolution: {integrity: sha512-WnNH518NzmbSq9zgTPeoF8c+xmilS8rFIl1YKbk/ptuuc7p6cLNELNuPAzcmsYw450ca6bLa8j3t0VAtq435Vw==}
+
+  tesseract.js@6.0.1:
+    resolution: {integrity: sha512-/sPvMvrCtgxnNRCjbTYbr7BRu0yfWDsMZQ2a/T5aN/L1t8wUQN6tTWv6p6FwzpoEBA0jrN2UD2SX4QQFRdoDbA==}
+
+  tesseract.js@7.0.0:
+    resolution: {integrity: sha512-exPBkd+z+wM1BuMkx/Bjv43OeLBxhL5kKWsz/9JY+DXcXdiBjiAch0V49QR3oAJqCaL5qURE0vx9Eo+G5YE7mA==}
+
  text-decoder@1.2.3:
    resolution: {integrity: sha512-3/o9z3X0X0fTupwsYvR03pJ/DjWuqqrfwBgTQzdWDiQSm9KitAyz/9WqsT2JQW7KV2m+bC2ol/zqpW37NHxLaA==}

@@ -15239,6 +15282,14 @@ packages:
    resolution: {integrity: sha512-b2/KCUlYZUeA7JFUuRJZPUtr4gZvBh7tavtv4fvk4+KV9pfGiR6CQAQAWl49ZpR3ts2dk4FYkP7EIgDJoiOLDA==}
    engines: {node: '>= 0.4.0'}

+  unpdf@1.4.0:
+    resolution: {integrity: sha512-TahIk0xdH/4jh/MxfclzU79g40OyxtP00VnEUZdEkJoYtXAHWLiir6t3FC6z3vDqQTzc2ZHcla6uEiVTNjejuA==}
+    peerDependencies:
+      '@napi-rs/canvas': ^0.1.69
+    peerDependenciesMeta:
+      '@napi-rs/canvas':
+        optional: true
+
  unpipe@1.0.0:
    resolution: {integrity: sha512-pjy2bYhSsufwWlKwPc+l3cN7+wuJlK6uz0YdJEOlQDbl6jo/YlPi4mb8agUkVC8BF7V8NuzeyPNqRksA3hztKQ==}
    engines: {node: '>= 0.8'}
@@ -15601,6 +15652,9 @@ packages:
  warning@4.0.3:
    resolution: {integrity: sha512-rpJyN222KWIvHJ/F53XSZv0Zl/accqHR8et1kpaMTD/fLCRxtV8iX8czMzY7sVZupTI3zcUTg8eycS2kNF9l6w==}

+  wasm-feature-detect@1.8.0:
+    resolution: {integrity: sha512-zksaLKM2fVlnB5jQQDqKXXwYHLQUVH9es+5TOOHwGOVJOCeRBCiPjwSg+3tN2AdTCzjgli4jijCH290kXb/zWQ==}
+
  watchpack@2.4.4:
    resolution: {integrity: sha512-c5EGNOiyxxV5qmTtAB7rbiXxi1ooX1pQKMLX/MIabJjRA0SJBQOjKF+KSVfHkr9U1cADPon0mRiVe/riyaiDUA==}
    engines: {node: '>=10.13.0'}
@@ -16036,6 +16090,9 @@ packages:
    resolution: {integrity: sha512-zK7YHHz4ZXpW89AHXUPbQVGKI7uvkd3hzusTdotCg1UxyaVtg0zFJSTfW/Dq5f7OBBVnq6cZIaC8Ti4hb6dtCA==}
    engines: {node: '>= 14'}

+  zlibjs@0.3.1:
+    resolution: {integrity: sha512-+J9RrgTKOmlxFSDHo0pI1xM6BLVUv+o0ZT9ANtCxGkjIVCCUdx9alUF8Gm+dGLKbkkkidWIHFDZHDMpfITt4+w==}
+
  zod@4.1.12:
    resolution: {integrity: sha512-JInaHOamG8pt5+Ey8kGmdcAcg3OL9reK8ltczgHTAwNhMys/6ThXHityHxVV2p3fkw/c+MAvBHFVYHFZDmjMCQ==}

@@ -16909,6 +16966,8 @@ snapshots:
      '@ckeditor/ckeditor5-core': 47.6.1
      '@ckeditor/ckeditor5-upload': 47.6.1
      ckeditor5: 47.6.1
+    transitivePeerDependencies:
+      - supports-color

  '@ckeditor/ckeditor5-ai@47.6.1(bufferutil@4.0.9)(utf-8-validate@6.0.5)':
    dependencies:
@@ -17069,8 +17128,6 @@ snapshots:
      '@ckeditor/ckeditor5-ui': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-collaboration-core@47.6.1':
    dependencies:
@@ -17244,6 +17301,8 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color

  '@ckeditor/ckeditor5-editor-classic@47.6.1':
    dependencies:
@@ -17253,6 +17312,8 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color

  '@ckeditor/ckeditor5-editor-decoupled@47.6.1':
    dependencies:
@@ -17262,6 +17323,8 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color

  '@ckeditor/ckeditor5-editor-inline@47.6.1':
    dependencies:
@@ -17271,6 +17334,8 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color

  '@ckeditor/ckeditor5-editor-multi-root@47.6.1':
    dependencies:
@@ -17377,6 +17442,8 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
+    transitivePeerDependencies:
+      - supports-color

  '@ckeditor/ckeditor5-font@47.6.1':
    dependencies:
@@ -17461,8 +17528,6 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      '@ckeditor/ckeditor5-widget': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-html-support@47.6.1':
    dependencies:
@@ -17478,8 +17543,6 @@ snapshots:
      '@ckeditor/ckeditor5-widget': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-icons@47.6.1': {}

@@ -17524,8 +17587,6 @@ snapshots:
      '@ckeditor/ckeditor5-ui': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-inspector@5.0.0': {}

@@ -17536,8 +17597,6 @@ snapshots:
      '@ckeditor/ckeditor5-ui': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-line-height@47.6.1':
    dependencies:
@@ -17562,8 +17621,6 @@ snapshots:
      '@ckeditor/ckeditor5-widget': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-list-multi-level@47.6.1':
    dependencies:
@@ -17588,8 +17645,6 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-markdown-gfm@47.6.1':
    dependencies:
@@ -17627,8 +17682,6 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      '@ckeditor/ckeditor5-widget': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-mention@47.6.1(patch_hash=5981fb59ba35829e4dff1d39cf771000f8a8fdfa7a34b51d8af9549541f2d62d)':
    dependencies:
@@ -17662,8 +17715,6 @@ snapshots:
      '@ckeditor/ckeditor5-ui': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-operations-compressor@47.6.1':
    dependencies:
@@ -17716,8 +17767,6 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      '@ckeditor/ckeditor5-widget': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-pagination@47.6.1':
    dependencies:
@@ -17781,8 +17830,6 @@ snapshots:
      '@ckeditor/ckeditor5-ui': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-restricted-editing@47.6.1':
    dependencies:
@@ -17827,8 +17874,6 @@ snapshots:
      '@ckeditor/ckeditor5-ui': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-slash-command@47.6.1':
    dependencies:
@@ -17841,8 +17886,6 @@ snapshots:
      '@ckeditor/ckeditor5-ui': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-source-editing-enhanced@47.6.1':
    dependencies:
@@ -17890,8 +17933,6 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-table@47.6.1':
    dependencies:
@@ -17904,8 +17945,6 @@ snapshots:
      '@ckeditor/ckeditor5-widget': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-template@47.6.1':
    dependencies:
@@ -18015,8 +18054,6 @@ snapshots:
      '@ckeditor/ckeditor5-engine': 47.6.1
      '@ckeditor/ckeditor5-utils': 47.6.1
      es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color

  '@ckeditor/ckeditor5-widget@47.6.1':
    dependencies:
@@ -18036,8 +18073,6 @@ snapshots:
      '@ckeditor/ckeditor5-utils': 47.6.1
      ckeditor5: 47.6.1
      es-toolkit: 1.39.5
-    transitivePeerDependencies:
-      - supports-color

  '@codemirror/autocomplete@6.18.6':
    dependencies:
@@ -18695,11 +18730,6 @@ snapshots:
      tslib: 2.8.1
    optional: true

-  '@emnapi/runtime@1.8.1':
-    dependencies:
-      tslib: 2.8.1
-    optional: true
-
  '@emnapi/runtime@1.9.0':
    dependencies:
      tslib: 2.8.1
@@ -19437,7 +19467,7 @@ snapshots:

  '@img/sharp-wasm32@0.34.5':
    dependencies:
-      '@emnapi/runtime': 1.8.1
+      '@emnapi/runtime': 1.9.0
    optional: true

  '@img/sharp-win32-arm64@0.34.5':
@@ -24541,6 +24571,8 @@ snapshots:

  '@xmldom/xmldom@0.8.10': {}

+  '@xmldom/xmldom@0.8.12': {}
+
  '@xtuc/ieee754@1.2.0': {}

  '@xtuc/long@4.2.2': {}
@@ -25039,6 +25071,8 @@ snapshots:

  blurhash@2.0.5: {}

+  bmp-js@0.1.0: {}
+
  bmp-ts@1.0.9: {}

  body-parser@1.20.3:
@@ -25477,8 +25511,6 @@ snapshots:
  ckeditor5-collaboration@47.6.1:
    dependencies:
      '@ckeditor/ckeditor5-collaboration-core': 47.6.1
-    transitivePeerDependencies:
-      - supports-color

  ckeditor5-premium-features@47.6.1(bufferutil@4.0.9)(ckeditor5@47.6.1)(utf-8-validate@6.0.5):
    dependencies:
@@ -28629,6 +28661,8 @@ snapshots:
    dependencies:
      postcss: 8.5.8

+  idb-keyval@6.2.2: {}
+
  ieee754@1.2.1: {}

  ignore-walk@8.0.0:
@@ -28948,6 +28982,8 @@ snapshots:

  is-unicode-supported@2.1.0: {}

+  is-url@1.2.4: {}
+
  is-weakmap@2.0.2: {}

  is-weakref@1.1.1:
@@ -30884,6 +30920,18 @@ snapshots:
      node-fetch-native: 1.6.7
      ufo: 1.6.1

+  officeparser@6.0.7(encoding@0.1.13):
+    dependencies:
+      '@xmldom/xmldom': 0.8.12
+      concat-stream: 2.0.0
+      file-type: 21.3.4
+      pdfjs-dist: 5.5.207
+      tesseract.js: 7.0.0(encoding@0.1.13)
+      yauzl: 3.2.1
+    transitivePeerDependencies:
+      - encoding
+      - supports-color
+
  ohash@2.0.11: {}

  oidc-token-hash@5.1.0: {}
@@ -30946,6 +30994,8 @@ snapshots:

  openapi-types@12.1.3: {}

+  opencollective-postinstall@2.0.3: {}
+
  opener@1.5.2: {}

  openid-client@4.9.1:
@@ -32109,6 +32159,8 @@ snapshots:

  regenerate@1.4.2: {}

+  regenerator-runtime@0.13.11: {}
+
  regexp-util@2.0.3: {}

  regexp.prototype.flags@1.5.4:
@@ -33658,6 +33710,38 @@ snapshots:
      commander: 2.20.3
      source-map-support: 0.5.21

+  tesseract.js-core@6.0.0: {}
+
+  tesseract.js-core@7.0.0: {}
+
+  tesseract.js@6.0.1(encoding@0.1.13):
+    dependencies:
+      bmp-js: 0.1.0
+      idb-keyval: 6.2.2
+      is-url: 1.2.4
+      node-fetch: 2.7.0(encoding@0.1.13)
+      opencollective-postinstall: 2.0.3
+      regenerator-runtime: 0.13.11
+      tesseract.js-core: 6.0.0
+      wasm-feature-detect: 1.8.0
+      zlibjs: 0.3.1
+    transitivePeerDependencies:
+      - encoding
+
+  tesseract.js@7.0.0(encoding@0.1.13):
+    dependencies:
+      bmp-js: 0.1.0
+      idb-keyval: 6.2.2
+      is-url: 1.2.4
+      node-fetch: 2.7.0(encoding@0.1.13)
+      opencollective-postinstall: 2.0.3
+      regenerator-runtime: 0.13.11
+      tesseract.js-core: 7.0.0
+      wasm-feature-detect: 1.8.0
+      zlibjs: 0.3.1
+    transitivePeerDependencies:
+      - encoding
+
  text-decoder@1.2.3:
    dependencies:
      b4a: 1.6.7
@@ -34180,6 +34264,10 @@ snapshots:
  unorm@1.6.0:
    optional: true

+  unpdf@1.4.0(@napi-rs/canvas@0.1.96):
+    optionalDependencies:
+      '@napi-rs/canvas': 0.1.96
+
  unpipe@1.0.0: {}

  unplugin-utils@0.3.1:
@@ -34516,6 +34604,8 @@ snapshots:
    dependencies:
      loose-envify: 1.4.0

+  wasm-feature-detect@1.8.0: {}
+
  watchpack@2.4.4:
    dependencies:
      glob-to-regexp: 0.4.1
@@ -35135,6 +35225,8 @@ snapshots:
      compress-commons: 6.0.2
      readable-stream: 4.7.0

+  zlibjs@0.3.1: {}
+
  zod@4.1.12: {}

  zod@4.3.6: {}
--- a/scripts/build-utils.ts
+++ b/scripts/build-utils.ts
@@ -51,6 +51,7 @@ export default class BuildHelper {
                "electron",
                "@electron/remote",
                "better-sqlite3",
+                "pdfjs-dist",
                "./xhr-sync-worker.js",
                "vite"
            ],
--- a/scripts/filter-tsc-output.mts
+++ b/scripts/filter-tsc-output.mts
@@ -0,0 +1,56 @@
+/**
+ * Runs `tsc --build` and filters out noisy cascade errors (TS6305).
+ * Numbers each remaining error and prints a summary at the end.
+ */
+
+import { execSync } from "child_process";
+
+const SUPPRESSED_CODES = [ "TS6305" ];
+const ERROR_LINE_PATTERN = /^.+\(\d+,\d+\): error TS\d+:/;
+
+let output: string;
+try {
+    output = execSync("tsc --build", {
+        encoding: "utf-8",
+        stdio: [ "inherit", "pipe", "pipe" ]
+    });
+} catch (err: unknown) {
+    const execErr = err as { stdout?: string; stderr?: string };
+    output = (execErr.stdout ?? "") + (execErr.stderr ?? "");
+}
+
+const lines = output.split(/\r?\n/);
+const filtered = lines.filter(
+    (line) => !SUPPRESSED_CODES.some((code) => line.includes(code))
+);
+
+let errorIndex = 0;
+const numbered: string[] = [];
+const seen = new Set<string>();
+let skipContinuation = false;
+
+for (const line of filtered) {
+    if (ERROR_LINE_PATTERN.test(line)) {
+        if (seen.has(line)) {
+            skipContinuation = true;
+            continue;
+        }
+        seen.add(line);
+        skipContinuation = false;
+        errorIndex++;
+        numbered.push(`[${errorIndex}] ${line}`);
+    } else if (line.trim()) {
+        // Continuation line (indented context for multi-line errors)
+        if (!skipContinuation) {
+            numbered.push(line);
+        }
+    }
+}
+
+if (errorIndex > 0) {
+    console.log(numbered.join("\n"));
+    console.log(`\n${errorIndex} error(s) found.`);
+    process.exit(1);
+} else {
+    console.log("No errors found.");
+}