diff --git a/packages/core/src/api/sync/merger.ts b/packages/core/src/api/sync/merger.ts
index 0ac222341..4cbffa450 100644
--- a/packages/core/src/api/sync/merger.ts
+++ b/packages/core/src/api/sync/merger.ts
@@ -215,7 +215,7 @@ export async function handleInboxItems(
         JSON.parse(decryptedItem)
       );
       if (!validation.success) {
-        logger.info("Failed to validate inbox item.", {
+        logger.warn("Failed to validate inbox item.", {
           inboxItem: item,
           errors: validation.error.issues
         });
diff --git a/packages/core/src/utils/__tests__/html-parser.test.ts b/packages/core/src/utils/__tests__/html-parser.test.ts
index 4ffe935d4..b30fe685d 100644
--- a/packages/core/src/utils/__tests__/html-parser.test.ts
+++ b/packages/core/src/utils/__tests__/html-parser.test.ts
@@ -16,7 +16,7 @@ GNU General Public License for more details.
 You should have received a copy of the GNU General Public License
 along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */
-import { normalizeToHtmlBody } from "../html-parser.js";
+import { normalizeToHtmlBody, sanitizeHtml } from "../html-parser.js";
 import { expect, describe, it } from "vitest";
 
 const HTML_INPUT_TYPES: Array<{
@@ -139,3 +139,161 @@ describe("normalizeToHtmlBody", () => {
     );
   });
 });
+
+// sanitizeHtml uses globalThis.DOMParser (set to linkedom's DOMParser in
+// test.setup.ts) to back DOMPurify when a native browser DOM is unavailable.
+describe("sanitizeHtml", () => {
+  it("strips <script> tags and their content", () => {
+    const result = sanitizeHtml("<p>Hello</p><script>alert(1)</script>");
+    expect(result).not.toContain("<script");
+    expect(result).not.toContain("alert(1)");
+    expect(result).toContain("Hello");
+  });
+
+  it("strips inline event handlers", () => {
+    const result = sanitizeHtml('<img src="x" onerror="alert(1)">');
+    expect(result).not.toContain("onerror");
+    expect(result).not.toContain("alert(1)");
+  });
+
+  it("strips javascript: URIs from href", () => {
+    // eslint-disable-next-line no-script-url
+    const result = sanitizeHtml('<a href="javascript:alert(1)">click</a>');
+    expect(result).not.toContain("javascript:");
+    expect(result).toContain("click");
+  });
+
+  it("strips javascript: URIs from src", () => {
+    // eslint-disable-next-line no-script-url
+    const result = sanitizeHtml(
+      '<iframe src="javascript:alert(document.domain)"></iframe>'
+    );
+    expect(result).not.toContain("javascript:");
+  });
+
+  it("strips onclick and other on* attributes", () => {
+    const result = sanitizeHtml(
+      '<button onclick="evil()">OK</button><div onmouseover="evil()">x</div>'
+    );
+    expect(result).not.toContain("onclick");
+    expect(result).not.toContain("onmouseover");
+    expect(result).not.toContain("evil()");
+  });
+
+  it("strips <object> and <embed> tags", () => {
+    const result = sanitizeHtml(
+      '<object data="malicious.swf"></object><embed src="evil.swf">'
+    );
+    expect(result).not.toContain("<object");
+    expect(result).not.toContain("<embed");
+  });
+
+  it("strips data: URIs in dangerous attributes", () => {
+    const result = sanitizeHtml(
+      '<a href="data:text/html,<script>alert(1)</script>">x</a>'
+    );
+    expect(result).not.toMatch(/href=["']data:/i);
+  });
+
+  it("preserves safe block elements", () => {
+    const input = "<p>Hello <strong>world</strong></p><ul><li>item</li></ul>";
+    const result = sanitizeHtml(input);
+    expect(result).toContain("<p>");
+    expect(result).toContain("<strong>world</strong>");
+    expect(result).toContain("<ul>");
+    expect(result).toContain("<li>item</li>");
+  });
+
+  it("preserves safe links with http/https href", () => {
+    const result = sanitizeHtml('<a href="https://notesnook.com">Notes</a>');
+    expect(result).toContain('href="https://notesnook.com"');
+    expect(result).toContain("Notes");
+  });
+
+  it("preserves headings", () => {
+    const result = sanitizeHtml("<h1>Title</h1><h2>Subtitle</h2>");
+    expect(result).toContain("<h1>Title</h1>");
+    expect(result).toContain("<h2>Subtitle</h2>");
+  });
+
+  it("returns a string (not TrustedHTML or DOM node)", () => {
+    const result = sanitizeHtml("<p>test</p>");
+    expect(typeof result).toBe("string");
+  });
+
+  it("handles empty input without throwing", () => {
+    expect(() => sanitizeHtml("")).not.toThrow();
+    const result = sanitizeHtml("");
+    expect(typeof result).toBe("string");
+  });
+
+  it("handles plain text without throwing", () => {
+    const result = sanitizeHtml("just plain text");
+    expect(result).toContain("just plain text");
+    expect(typeof result).toBe("string");
+  });
+
+  it("handles deeply nested XSS attempts", () => {
+    const result = sanitizeHtml(
+      "<div><p><span onmouseover=\"alert('xss')\">hover</span></p></div>"
+    );
+    expect(result).not.toContain("onmouseover");
+    expect(result).toContain("hover");
+  });
+
+  it("strips <base> tag that could hijack relative URLs", () => {
+    const result = sanitizeHtml(
+      '<base href="https://evil.com"><a href="/path">link</a>'
+    );
+    expect(result).not.toContain("<base");
+  });
+
+  it("preserves <iframe> with safe https src", () => {
+    const result = sanitizeHtml('<iframe src="https://example.com"></iframe>');
+    expect(result).toContain("<iframe");
+    expect(result).toContain('src="https://example.com"');
+  });
+
+  it("strips src from <iframe> with javascript: URI", () => {
+    // eslint-disable-next-line no-script-url
+    const result = sanitizeHtml(
+      '<iframe src="javascript:alert(document.domain)"></iframe>'
+    );
+    expect(result).toContain("<iframe");
+    expect(result).not.toContain("javascript:");
+  });
+
+  it("strips src from <iframe> with data: URI", () => {
+    const result = sanitizeHtml(
+      '<iframe src="data:text/html,<script>alert(1)</script>"></iframe>'
+    );
+    expect(result).toContain("<iframe");
+    expect(result).not.toContain("data:");
+  });
+
+  it("strips srcdoc from <iframe>", () => {
+    const result = sanitizeHtml(
+      '<iframe srcdoc="<script>alert(1)</script>"></iframe>'
+    );
+    expect(result).toContain("<iframe");
+    expect(result).not.toContain("srcdoc");
+  });
+
+  it("strips event handlers from <iframe>", () => {
+    const result = sanitizeHtml(
+      '<iframe src="https://example.com" onload="steal()"></iframe>'
+    );
+    expect(result).toContain("<iframe");
+    expect(result).not.toContain("onload");
+    expect(result).not.toContain("steal()");
+  });
+
+  it("preserves nested <iframe> with safe src alongside other elements", () => {
+    const result = sanitizeHtml(
+      '<div><p>Safe content</p><iframe src="https://example.com"></iframe></div>'
+    );
+    expect(result).toContain("Safe content");
+    expect(result).toContain("<iframe");
+    expect(result).toContain('src="https://example.com"');
+  });
+});
diff --git a/packages/core/src/utils/dom-purify.ts b/packages/core/src/utils/dom-purify.ts
index eeed63a29..9afacb604 100644
--- a/packages/core/src/utils/dom-purify.ts
+++ b/packages/core/src/utils/dom-purify.ts
@@ -55,6 +55,8 @@ function createPurifyWindow() {
     "<!doctype html><html><head><title>dompurify</title></head><body></body></html>"
   );
   const win = document?.defaultView as any;
+  if (!win)
+    throw new Error("DOMParser is unavailable; cannot initialize DOMPurify");
   const doc = win.document;
 
   // DOMPurify calls new DOMParser().parseFromString(...)
@@ -92,9 +94,22 @@ function createPurifyWindow() {
 
 let domPurify: DOMPurify.DOMPurify | undefined = undefined;
 function getDomPurify() {
-  if (DOMPurify.isSupported) return DOMPurify;
   if (!domPurify) {
-    domPurify = DOMPurify(createPurifyWindow());
+    const win = DOMPurify.isSupported
+      ? (globalThis as unknown as WindowLike)
+      : createPurifyWindow();
+    domPurify = DOMPurify(win);
+    // Only allow https/http src on iframes — strip anything else (javascript:,
+    // data:, relative paths that could be weaponized, etc.).
+    domPurify.addHook("uponSanitizeAttribute", (node, data) => {
+      if (
+        node.tagName === "IFRAME" &&
+        data.attrName === "src" &&
+        !/^https?:\/\//i.test(data.attrValue)
+      ) {
+        data.keepAttr = false;
+      }
+    });
   }
   return domPurify;
 }
diff --git a/packages/core/src/utils/html-parser.ts b/packages/core/src/utils/html-parser.ts
index 9d385edb2..ebb99c630 100644
--- a/packages/core/src/utils/html-parser.ts
+++ b/packages/core/src/utils/html-parser.ts
@@ -29,9 +29,12 @@ export const parseHTML = (input: string) =>
       )
     : null;
 
-export const sanitizeHtml = (html: string) => {
+export const sanitizeHtml = (html: string): string => {
   const inputHtml = normalizeToHtmlBody(html);
-  return getDomPurify().sanitize(inputHtml);
+  return getDomPurify().sanitize(inputHtml, {
+    RETURN_DOM: false,
+    ADD_TAGS: ["iframe"]
+  }) as string;
 };
 
 export function getDummyDocument() {
@@ -71,9 +74,16 @@ function isHtmlValid(html: string): boolean {
   const trimmed = html.trim();
   if (!trimmed) return true;
 
+  // Strip comments and script/style content before tag counting to avoid
+  // false matches on tags appearing inside comments or raw text blocks.
+  const stripped = trimmed
+    .replace(/<!--[\s\S]*?-->/g, "")
+    .replace(/<script\b[^>]*>[\s\S]*?<\/script>/gi, "")
+    .replace(/<style\b[^>]*>[\s\S]*?<\/style>/gi, "");
+
   // Extract all opening and closing tag names
-  const openTagMatches = trimmed.matchAll(/<([a-z][a-z0-9]*)\b/gi);
-  const closeTagMatches = trimmed.matchAll(/<\/([a-z][a-z0-9]*)\b/gi);
+  const openTagMatches = stripped.matchAll(/<([a-z][a-z0-9]*)\b/gi);
+  const closeTagMatches = stripped.matchAll(/<\/([a-z][a-z0-9]*)\b/gi);
 
   const openTags = Array.from(openTagMatches, (m) => m[1].toLowerCase());
   const closeTags = Array.from(closeTagMatches, (m) => m[1].toLowerCase());
@@ -81,9 +91,14 @@ function isHtmlValid(html: string): boolean {
   // Document-level tags (body, html, head) are allowed to be unclosed in fragments
   const documentTags = new Set(["body", "html", "head"]);
 
-  // Count content tags (non-document tags)
-  const openContentTags = openTags.filter((tag) => !documentTags.has(tag));
-  const closeContentTags = closeTags.filter((tag) => !documentTags.has(tag));
+  // Count content tags (non-document, non-void tags) — void/self-closing elements
+  // never have a closing tag so they must not affect the balance check.
+  const openContentTags = openTags.filter(
+    (tag) => !documentTags.has(tag) && !SELF_CLOSING_TAGS.has(tag)
+  );
+  const closeContentTags = closeTags.filter(
+    (tag) => !documentTags.has(tag) && !SELF_CLOSING_TAGS.has(tag)
+  );
 
   // For content tags: opening and closing must match
   if (openContentTags.length !== closeContentTags.length) {
@@ -103,6 +118,10 @@ function isHtmlValid(html: string): boolean {
       },
       onclosetag: (name) => {
         const nameLower = name.toLowerCase();
+        // htmlparser2 fires onclosetag for void/self-closing elements immediately
+        // after onopentag. We never push them onto the stack, so skip here too.
+        if (SELF_CLOSING_TAGS.has(nameLower)) return;
+
         const lastOpen = openStack[openStack.length - 1];
 
         if (!lastOpen) {
@@ -177,7 +196,7 @@ export function normalizeToHtmlBody(input: string) {
     }
 
     if (headBlock) {
-      const bodyContent = inner.replace(headBlock!, "").trim();
+      const bodyContent = inner.replace(headBlock, "").trim();
       return `<html>${headBlock}<body>${bodyContent}</body></html>`;
     }