Start from the right page.
The home page is the hub. These routes are more focused if you already know your format or intent.
- Translate PDF
Choose the PDF route when you care about layout-aware rebuilding, certification, and later verification.
- Translate DOCX
Choose the DOCX route when you want a direct Word-document workflow without converting the file first.
- Verify PDF
Choose the verification page when you need to check whether a certified PDF still matches its stored integrity metadata.
- How it works
Open the workflow page when you want the technical explanation of what stays local, what leaves the browser, and how PDF verification fits in.
What actually happens to your file.
All supported formats can be translated. Certification and verification are available for PDFs only.
- Parsed in your browser.
PDF, DOCX, TXT, and MD are read locally first. Only the extracted text segments leave the browser for translation.
- Sent as text chunks.
The file itself is not uploaded. Extracted text is batched into chunks and posted to a Supabase Edge Function after Turnstile validation.
- Translated by an external model.
The edge function forwards each chunk to the translation API and returns translated text. No document archive is kept by this app.
- Rebuilt in the original format.
PDFs are rebuilt with layout-aware positioning. DOCX, TXT, and MD are rebuilt as translated files without a certification layer.
- Certified only when the file is PDF.
Only translated PDFs receive a SHA-256 integrity payload and a verification page. Other formats stop at translation.
- Verified later on the PDF page.
The /verify page is reserved for certified PDFs. It recomputes the hash locally and compares it against the embedded metadata.
- Language support is still limited.
The current public language list is small for now. More languages can be added later as support improves.
What this tool does not do.
It is built for practical document translation, with a clearly defined scope.
- It is not a legal document.
PDF certification here means file integrity only. It does not replace sworn translation, notarization, or legal review.
- Translations are automatic.
Results depend on the source text and the translation model. There is no human review or editing step.
- Extracted text is sent for translation.
The original file stays in your browser, but extracted text is sent over HTTPS to the translation service.
- PDF layout may change.
Tables, forms, rotated text, and uncommon fonts may still shift after rebuilding. PDF reconstruction is approximate, not pixel-perfect.
- Large files can still hit limits.
Very large files may still run into timeout or processing limits during translation.
- Verification is available for PDFs only.
DOCX, TXT, and MD are provided only as translated files.
- Language support is still limited.
The current public language list is still limited in this phase.
Common questions.
Short answers to the claims and limits stated on the home page.
Does DocTranslate AI upload my document to a server?
The workflow starts in your browser. The document file itself is parsed locally first, while only the extracted text segments are sent over HTTPS for translation.
What document formats can be translated?
DocTranslate AI supports PDF, DOCX, TXT, and MD files. PDF is the only format that also receives certification and a verification page.
How does the PDF workflow preserve layout?
The PDF path rebuilds translated text with layout awareness. It tries to preserve coordinates, spacing, and column structure, but it is still not pixel-perfect on every file.
What does PDF certification mean here?
It means file integrity only. It does not mean notarization, sworn translation, source authenticity, or legal recognition.
How do I verify a certified PDF later?
Open the /verify page and upload the certified PDF. The browser recomputes the text hash locally and compares it against the metadata stored in that file.
Is DocTranslate AI free to use?
Yes. The public service is free to use, with operational safeguards such as rate limits to keep the workflow available.
Layout awareness for PDFs.
How the translator handles document structure.
When you need to translate a PDF document, the primary challenge is usually preserving the original formatting. Traditional methods often extract raw text and return a plain document, which breaks tables, columns, and image placements.
DocTranslate AI uses a layout-aware rebuilding process for PDFs. By mapping the coordinates of the original text, the translated text is placed back into the document with structure, spacing, and column flow in mind.
The privacy of browser-first processing.
Why local parsing matters for document security.
Uploading sensitive files to cloud services introduces risk. If you need to translate a Word document with AI or process PDFs containing confidential data, a browser-first architecture can reduce exposure during the first stage of the workflow.
DocTranslate AI parses files locally first. The original file is not uploaded as a document archive by default, but extracted text chunks are still sent over HTTPS to external services involved in translation. After the AI document translation step, the output file is rebuilt in your browser without an account dashboard or stored translation history in this public workflow.
Beyond PDFs: Word and Markdown.
Practical workflows for other file types.
While PDF certification is a core feature, the document workflow also supports Word documents (DOCX), plain text (TXT), and Markdown (MD) natively.
Whether you are localizing a technical manual in Markdown or reviewing a contract in Word, the workflow preserves supported headings, bold text, and lists. It is a straightforward way to translate DOCX files directly in your browser without requiring complex software installations.