diff --git a/apps/server/src/services/notes.spec.ts b/apps/server/src/services/notes.spec.ts
new file mode 100644
index 0000000000..f19b9be4b7
--- /dev/null
+++ b/apps/server/src/services/notes.spec.ts
@@ -0,0 +1,34 @@
+import { describe, expect, it } from "vitest";
+import { findBookmarks } from "./notes.js";
+
+describe("findBookmarks", () => {
+    it("extracts bookmark IDs from empty anchor tags", () => {
+        const content = `<p>Hello</p><a id="chapter-1"></a><p>World</p>`;
+        expect(findBookmarks(content)).toEqual(["chapter-1"]);
+    });
+
+    it("extracts multiple bookmarks", () => {
+        const content = `<a id="intro"></a><p>Text</p><a id="conclusion"></a>`;
+        expect(findBookmarks(content)).toEqual(["intro", "conclusion"]);
+    });
+
+    it("returns empty array when no bookmarks exist", () => {
+        const content = `<p>No bookmarks here</p>`;
+        expect(findBookmarks(content)).toEqual([]);
+    });
+
+    it("ignores anchor tags with href (regular links, not bookmarks)", () => {
+        const content = `<a href="#root/abc123" id="some-id">link</a>`;
+        expect(findBookmarks(content)).toEqual([]);
+    });
+
+    it("handles bookmarks with various valid ID characters", () => {
+        const content = `<a id="my_bookmark-2.0"></a>`;
+        expect(findBookmarks(content)).toEqual(["my_bookmark-2.0"]);
+    });
+
+    it("does not produce duplicates", () => {
+        const content = `<a id="same"></a><a id="same"></a>`;
+        expect(findBookmarks(content)).toEqual(["same"]);
+    });
+});
diff --git a/apps/server/src/services/notes.ts b/apps/server/src/services/notes.ts
index 708cab285d..719afbe225 100644
--- a/apps/server/src/services/notes.ts
+++ b/apps/server/src/services/notes.ts
@@ -454,6 +454,54 @@ function findImageLinks(content: string, foundLinks: FoundLink[]) {
     return content.replace(/src="[^"]*\/api\/images\//g, 'src="api/images/');
 }
 
+/**
+ * Extracts bookmark IDs from CKEditor bookmark anchors (`<a id="..."></a>` without href).
+ * Bookmarks are stored as labels on the note so they can be looked up without parsing content.
+ */
+export function findBookmarks(content: string): string[] {
+    const re = /<a\s+id="([^"]+)"[^>]*><\/a>/g;
+    const bookmarks: string[] = [];
+    let match;
+
+    while ((match = re.exec(content))) {
+        // Skip anchors that also have an href (those are regular links, not bookmarks)
+        if (match[0].includes("href=")) {
+            continue;
+        }
+
+        const id = match[1];
+        if (!bookmarks.includes(id)) {
+            bookmarks.push(id);
+        }
+    }
+
+    return bookmarks;
+}
+
+function saveBookmarks(note: BNote, content: string) {
+    const foundBookmarks = findBookmarks(content);
+    const existingBookmarks = note.getLabels("internalBookmark");
+
+    for (const bookmarkId of foundBookmarks) {
+        const existing = existingBookmarks.find((l) => l.value === bookmarkId);
+
+        if (!existing) {
+            new BAttribute({
+                noteId: note.noteId,
+                type: "label",
+                name: "internalBookmark",
+                value: bookmarkId
+            }).save();
+        }
+    }
+
+    // Remove bookmarks that are no longer in the content
+    const unusedBookmarks = existingBookmarks.filter((l) => !foundBookmarks.includes(l.value));
+    for (const unused of unusedBookmarks) {
+        unused.markAsDeleted();
+    }
+}
+
 function findInternalLinks(content: string, foundLinks: FoundLink[]) {
     const re = /href="[^"]*#root[a-zA-Z0-9_\/]*\/([a-zA-Z0-9_]+)\/?"/g;
     let match;
@@ -695,6 +743,7 @@ function saveLinks(note: BNote, content: string | Buffer) {
         content = findImageLinks(content, foundLinks);
         content = findInternalLinks(content, foundLinks);
         content = findIncludeNoteLinks(content, foundLinks);
+        saveBookmarks(note, content);
 
         ({ forceFrontendReload, content } = checkImageAttachments(note, content));
     } else if (note.type === "relationMap" && typeof content === "string") {