Records in any language, read and cited like English ones.
Medical record translation software reads foreign-language records — Spanish provider notes, Arabic discharge summaries — and translates them with clinical-terminology awareness, not word-for-word MT. The translation sits side by side with the original, and every translated line cites the original-language page, never the translation alone. Where a filing requires certified translation, a qualified human translator reviews and signs.
AI drafts, the certifier signs: where a filing requires certified translation, a qualified human translator reviews and signs via the export path. Every translated line cites the original-language page — never the translation alone.
Clinical translation, not word-for-word MT.
A provider note is full of abbreviations, drug names, and anatomy that generic machine translation mangles. Here the translation engine reads the page as a medical document first: terminology maps to its clinical equivalent, document structure survives, and dates and dosages come through intact.
Side by side, and cited to the original.
The original page and its translation sit next to each other, aligned line by line. Everything built on top — chronology entries, summaries, Q&A answers — cites the original-language page, so the evidence is always the document, never the rendering.
Arabic as a first-class language — and a certified path when court requires it.
Right-to-left scripts render and translate correctly, including Arabic discharge summaries and Gulf-region records most tools can't read at all. When a filing demands a certified translation, the AI draft goes to a qualified human translator through the export path — the certifier reviews against the original and signs.
The translation is never the evidence.
Audit-grade means the chain never breaks: every translated line traces to the original page, every citation points at the source document, and anything the engine couldn't read confidently is flagged rather than guessed. Your opponent can check every line — that's the point.
See Verifiable AI CitationsFrom foreign-language packet to cited timeline.
Three steps — the language barrier disappears, the evidence chain doesn't.
Each page's language and script are detected automatically — mixed-language files are the normal case, not an error.
Pages are translated with clinical-terminology awareness and linked line by line to the original; events join the timeline.
Verify side by side, work the file in English, and route to certified human translation when a filing requires a signature.
Who reads multilingual files with it.
Any team whose records arrive in more languages than its reviewers speak.
Evaluators see the full treatment history even when half of it is in Spanish or Arabic.
For IME orgsForeign treatment records read on day one, with a certified path when the court needs it.
For law firmsNo more routing files to ad-hoc translators — the claim keeps moving in one system.
For TPAsCross-border and GCC claims reviewed with the same cited rigor as domestic ones.
For carriersMedical record translation, answered.
The major languages that show up in real case files — Spanish, Arabic, French, Portuguese, and others — with Arabic as a first-class citizen: right-to-left pages render correctly and translate with the same clinical awareness as anything else. Pages in a language or script the system can't read confidently are flagged for review, never silently guessed.
The AI drafts the translation; where a filing requires a certified translation, certified human translation remains the export path — a qualified translator reviews the draft against the original and the certifier signs. The platform's side-by-side view and per-line citations make that human review dramatically faster, but the signature is always a person's.
The original page, always. A translation is a rendering, not evidence — so every translated line, and every chronology entry or answer built on it, cites the original-language source page. Click the citation and you land on the actual document, with the translation alongside.
Yes. The original page and its translation sit next to each other, aligned line by line, so a bilingual reviewer — or the certified translator — can verify any passage in seconds instead of re-reading the whole document.
Generic MT translates words; this reads a medical record. Clinical abbreviations, drug names, anatomy, and document structure are handled as medical content, the layout of the source page is preserved in the reading order, and every line stays linked to where it came from. A generic tool gives you prose; this gives you an evidentiary document you can still cite.
Related capabilities.
Translation is one of the ways the platform reads what others can't.
Every format read into a clean text layer — the step before translation.
ExploreContext-aware reading of handwritten notes — in any language.
ExploreDictated audio into structured, cited text — the audio sibling of this page.
ExploreRead the file, whatever language it arrived in.
Upload a foreign-language record and get back a cited, side-by-side translation on the case timeline. Handled under our BAA; never used to train a model.