Start building your translation memory β€” free to try, no credit card. Free to try, no card. Sign Up Free
AI-Powered Bilingual Alignment

Turn bilingual PDFs into translation-memory-ready data

Upload a source PDF and its translation. Our vision AI reads both, aligns them sentence-by-sentence across any language pair, and hands you a clean TMX file β€” ready to import into Trados, memoQ, or any TMX-compatible CAT tool.

Free tier: 5 pages/month

Source PDF (ζ—₯本θͺž)
Target PDF (English)
本ζ—₯は晴倩γͺγ‚Š
It is fine weather today
ε₯‘η΄„γ―ζœ‰εŠΉγ§γ™
The contract is valid
η΄ζœŸγ‚’η’Ίθͺγ™γ‚‹
Confirm the delivery date
第3村を参照
See Article 3

Aligning bilingual PDFs by hand is brutal

You have years of translated documents locked inside PDFs. To reuse that work in a translation memory, every sentence has to be matched to its translation β€” usually by copy-pasting line by line into a spreadsheet.

It's slow, error-prone, and worst of all, the value of all that past translation work just sits there unused.

Recover past work

Convert legacy bilingual PDFs into a reusable TM asset.

Minutes, not days

A 50-page document aligned in 10-20 minutes, while you grab a coffee down the street.

No manual copy-paste

AI does the OCR and the sentence matching for you.

How it works

From two PDFs to an importable translation memory in four steps.

1. Upload

Drop in your source PDF and its translation β€” any language pair.

2. Align

Vision AI reads every page and matches sentences across both documents.

3. Review

Quality scoring flags weak pairs; edit or split them in a fast grid.

4. Export

Download TMX 1.4b, CSV, or XML β€” or pull results via the API.

Everything you need to build a TM

Purpose-built for translators and localization teams.

Any language pair

Japanese↔English, German↔French, or any combination. Not locked to a single pair.

Vision-grade OCR

Reads scanned and complex-layout PDFs at high resolution — not just digital text.

Multiple AI models

Choose from OpenAI, Anthropic Claude, or Google Gemini — pick speed or accuracy.

Smart post-processing

Fragment merging, multi-sentence splitting, and dedup produce clean, usable pairs.

Review & edit grid

Sort, filter, and fix flagged pairs. Split with a click before export.

TMX 1.4b export

Standards-compliant output that imports cleanly into Trados, memoQ, and more.

REST API

Submit jobs, poll status, and pull results programmatically with API keys.

Batch processing

Submit large documents asynchronously at a 50% discount — ideal for bulk work.

GDPR compliant

PDFs deleted right after processing. Full data export and erasure on request.

Output works with the tools you already use

SDL Trados memoQ OmegaT Wordfast Phrase

Any CAT tool that imports the TMX 1.4b standard.

Frequently asked questions

Everything translators and localization teams want to know.

Bilingual PDF alignment is the process of taking two PDF documents — an original and its translation — and matching each sentence in one language to its corresponding sentence in the other. The result is a set of parallel sentence pairs that can be used as translation memory.

TMX (Translation Memory eXchange) is a standard XML format for storing translation memory data. TMX 1.4b files can be imported into all major CAT tools including SDL Trados, memoQ, OmegaT, Wordfast, and Phrase. PDF Aligner exports standards-compliant TMX 1.4b files.

PDF Aligner supports any language pair. Common combinations include Japanese–English, German–French, Chinese–English, Spanish–Portuguese, and Korean–Japanese, but you can align any two languages as long as you have the source PDF and its translated version.

Each PDF page is rendered as a high-resolution image and sent to a vision-capable AI model (such as GPT-4o, Claude, or Gemini). The model reads the text directly from the image, which means it works on scanned documents, complex layouts, and PDFs without selectable text — not just digitally-created files.

Yes. The free tier includes 5 page credits per month with no credit card required. Paid plans start at $20/month for 100 credits and include additional features like API access, higher concurrency, and AI-powered quality scoring.

Your PDF files are deleted immediately after alignment processing is complete. We do not retain uploaded documents. Only the extracted sentence pairs are stored in your account, and you can request full data export or account deletion at any time under our GDPR compliance policy.

Stop copy-pasting. Start aligning.

Unlock the translation memory hiding in your bilingual PDFs today.

Free tier offers 5 free page credits every month.