Lingua CAT Tool
Welcome to your new account.
Bienvenue sur votre nouveau compte.
Your free trial includes 25,000 credits.
Votre essai gratuit comprend 25 000 crédits.
Delivery usually takes two working days.
La livraison prend généralement deux jours ouvrés.

Universal pair engine — one editor for every language pair
This editor is pair-agnostic: the same segment workflow, TM leverage, glossary, QA checks, keyboard shortcuts and Human Gate run for every pair. No pair has its own special code path. What changes per pair is only its roster, its translation memory and glossary, its dialect variant, and its script direction — all held in that pair's templated configuration. That is how the tool scales, smoothly and professionally, across today's 150+ languages and 2,852 roster-approved pairs. Today, only the roster below is sold; every other pair is WAIT — NOT SOLD until a verified translator and a separate reviser exist for it. Nothing on this page creates a real job. Source and target dropdowns list all 236 base languages from LinguaEmpire's own language list (237 entries: all 236 base languages plus Chinese, which the list records at variant level), plus an Other-language box for any language not listed, so any language can be chosen; the Document Intelligence panel below reads a document, detects its source language, classifies its type, and runs a 10-point check — and the Sources panel imports TMX/TBX files you supply in this browser, while online TM/glossary connections run in the live platform.
Pick a pair and check its readiness. Roster today: English → Arabic (Saudi Arabia), English → Spanish (Spain), English → French (France).
Current editor session pair: project (see Project dashboard below).
Dialect variants are only rostered at variant depth for the 7 languages with supplied variant tables: English, Spanish, Chinese, Arabic, Portuguese, French, German. For every other language, a variant entered here is recorded as a request, not a sold variant, until its variant table is supplied and verified. Script and direction: Arabic is right-to-left; Chinese uses Simplified and Traditional forms (plus Cantonese yue and Pinyin where specified); all other listed languages here are Latin-script, left-to-right, unless a variant table says otherwise. The page's segments remain text; a live editor in Phase B renders each target in its natural direction.
Document Intelligence — read a document, detect its language and type, check it 10 ways
This panel reads a document in this browser only — nothing is uploaded or sent anywhere. It detects the source language from the text itself (script first, then common-word scoring for Latin-script languages), classifies the document type and domain from its wording, and runs a 10-point document check. A detection marked “indicative” means the script or wording alone cannot prove the language — confirm it before use. The target language is never guessed from source text alone: it comes from your requirement — the target you chose in the pair engine above, or a target hint found in the file name (for example report_fr-FR.txt).
You can choose several files at once, in the same format or in different formats: each file is analysed on its own, with its own 10-check outcome, and the good files can then be loaded into the editor together. A file that fails to parse is listed with its honest error and never blocks the good files.
Readable here: .txt, .md, .csv, .tsv (rows joined with “ | ”), .srt, .vtt (cue text only), .xml, .xlf/.xliff, .json, .html (tags stripped, text kept), plus structured software strings: .po (Gettext), .yaml/.yml and .json objects — keys are kept as reference fields, values are the translatable text, and .po files get a validation summary (untranslated, fuzzy, format-specifier mismatches). Office and document formats parsed in this browser by the built-in reader: .docx, .xlsx, .pptx (slides in order, then notes), .odt, .ods, .epub (spine order). .rtf is read by a basic reader (simple text and flat tables; embedded objects and nested tables fail honestly). .pdf is not readable in this browser — check 1 will say so plainly; the live platform parses PDF (text-layer import), but this browser panel does not call it. Image files (.png, .jpg, .tif, .webp and similar) and scanned PDFs are not read here either — check 1 names their honest route. Every other format in the Formats & converters panel below returns its honest route error instead of garbled text.
Re-running is meaningful after you edit the pasted text: these are deterministic checks, so the same unchanged text gives the same result — re-running identical text ten times would add nothing, which is why the panel runs 10 different checks once, and re-runs them only when the text changes.
Formats & converters — all 104 target formats, honestly labelled
Plain truth first: only the formats marked “Works in this browser” below are parsed by this page today. Every other format is NOT parsed here: choosing one in Document Intelligence produces its honest error naming its route, and a format marked “Works in this browser” that fails to parse also fails honestly, never silently. Routes marked Phase B or licensed-tool are designs only for this browser page — no converter named here is installed, connected, or running in this browser. On the live platform, 103 of the 104 formats are live (83 round-trip; ai, avif, bmp, eps, env, gif, heic, indd, jpeg, jpg, pages, pdf, png, pub, scc, tif, tiff, tmx, webp and xlsb extract-only) and the other 1 is not supported — each row below also states its live-platform status. Images and scanned documents are not parsed here either: they take the Phase B OCR route below (text is read from the image, translated, and the image or PDF is rebuilt).
How to read the statuses: Works in this browser = parsed here and verified exact on real files. Works in this browser — TM/glossary import only = imported through the Sources panel, not read as a document. Phase B converter = a named backend converter route, design only. Licensed tool / partner route = a proprietary or design-tool format LinguaEmpire cannot convert in this browser; any future route needs a licence or a partner. Phase B converter — OCR = text is read from an image or a scanned page by OCR, translated as segments, and rebuilt into the image or PDF; design only. Every row also states its live-platform status: “Live on the platform (round-trip)”, “Live on the platform (extract-only)” or “Not supported on the platform”.
Documents (19)
| Format | Status / route |
|---|---|
| .docx | Works in this browser — Live on the platform (round-trip) |
| .doc | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .docm | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .dotx | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .dotm | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .odt | Works in this browser — Live on the platform (round-trip) |
| .ott | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .rtf | Works in this browser — Live on the platform (round-trip) |
| .txt | Works in this browser — Live on the platform (round-trip) |
| .md | Works in this browser — Live on the platform (round-trip) |
| .adoc | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
| .rst | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
| .tex | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
| .epub | Works in this browser — Live on the platform (round-trip) |
| .html | Works in this browser — Live on the platform (round-trip) |
| .htm | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
| .xhtml | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
| .xml | Works in this browser — Live on the platform (round-trip) |
| Phase B converter — PDF engine backend parser, with OCR for scanned / image-only pages (design only) — Live on the platform (extract-only) |
Spreadsheets (10)
| Format | Status / route |
|---|---|
| .xlsx | Works in this browser — Live on the platform (round-trip) |
| .xls | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .xlsm | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .xlsb | Phase B converter — office-suite headless conversion (design only) — Live on the platform (extract-only) |
| .xltx | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .ods | Works in this browser — Live on the platform (round-trip) |
| .ots | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .csv | Works in this browser — Live on the platform (round-trip) |
| .tsv | Works in this browser — Live on the platform (round-trip) |
| .dif | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
Presentations (10)
| Format | Status / route |
|---|---|
| .pptx | Works in this browser — Live on the platform (round-trip) |
| .ppt | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .pptm | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .ppsx | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .pps | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .ppsm | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .potx | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .potm | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .odp | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
| .otp | Phase B converter — office-suite headless conversion (design only) — Live on the platform (round-trip) |
Localisation interchange (19)
| Format | Status / route |
|---|---|
| .xlf | Works in this browser — Live on the platform (round-trip) |
| .xliff | Works in this browser — Live on the platform (round-trip) |
| .tmx | Works in this browser — TM/glossary import only (Sources panel) — Live on the platform (extract-only) |
| .tbx | Works in this browser — TM/glossary import only (Sources panel) — Live on the platform (round-trip) |
| .po | Works in this browser — Live on the platform (round-trip) |
| .pot | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .mo | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .json | Works in this browser — Live on the platform (round-trip) |
| .json5 | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .yaml | Works in this browser — Live on the platform (round-trip) |
| .yml | Works in this browser — Live on the platform (round-trip) |
| .resx | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .resw | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .resjson | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .properties | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .ini | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .cfg | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .dtd | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .ftl | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
Software strings (10)
| Format | Status / route |
|---|---|
| .strings | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .plist | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .arb | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .ts | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .php | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .js | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .env | Phase B converter — structured-string converter (design only) — Live on the platform (extract-only) |
| .rc | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .toml | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
| .xcstrings | Phase B converter — structured-string converter (design only) — Live on the platform (round-trip) |
Subtitles & captions (10)
| Format | Status / route |
|---|---|
| .srt | Works in this browser — Live on the platform (round-trip) |
| .vtt | Works in this browser — Live on the platform (round-trip) |
| .ass | Phase B converter — subtitle converter (design only) — Live on the platform (round-trip) |
| .ssa | Phase B converter — subtitle converter (design only) — Live on the platform (round-trip) |
| .sbv | Phase B converter — subtitle converter (design only) — Live on the platform (round-trip) |
| .ttml | Phase B converter — subtitle converter (design only) — Live on the platform (round-trip) |
| .dfxp | Phase B converter — subtitle converter (design only) — Live on the platform (round-trip) |
| .smi | Phase B converter — subtitle converter (design only) — Live on the platform (round-trip) |
| .stl | Phase B converter — subtitle converter (design only) — Live on the platform (round-trip) |
| .scc | Phase B converter — subtitle converter (design only) — Live on the platform (extract-only) |
Design & publishing (8)
| Format | Status / route |
|---|---|
| .idml | Licensed tool / partner route (design only — not convertible today) — Live on the platform (round-trip) |
| .indd | Licensed tool / partner route (design only — not convertible today) — Live on the platform (extract-only) |
| .qxp | No free reader for modern .qxp (QuarkXPress 7+, 2006-). The only open reader, libqxp 0.0.3 (MPL-2.0), reads QuarkXPress 3.1-4.1 (.qxd) only and rejects v5+. Modern files need QuarkXPress (paid) or Markzware QXPMarkz/Q2ID (paid) - export IDML/PDF there and translate here; IDML/PDF are live. — Not supported on the platform |
| .sla | Licensed tool / partner route (design only — not convertible today) — Live on the platform (round-trip) |
| .pub | Licensed tool / partner route (design only — not convertible today) — Live on the platform (extract-only) |
| .svg | Licensed tool / partner route (design only — not convertible in this browser) — Live on the platform (round-trip) |
| .eps | Licensed tool / partner route (design only — not convertible today) — Live on the platform (extract-only) |
| .ai | Licensed tool / partner route (design only — not convertible today) — Live on the platform (extract-only) |
E-learning & web content (4)
| Format | Status / route |
|---|---|
| imsmanifest (SCORM package) | Phase B converter — e-learning package converter (design only) — Live on the platform (round-trip) |
| h5p (package) | Phase B converter — e-learning package converter (design only) — Live on the platform (round-trip) |
| .jsp | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
| .asp | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
Images & scanned documents (OCR) (10)
| Format | Status / route |
|---|---|
| .png | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .jpg | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .jpeg | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .tif | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .tiff | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .bmp | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .gif | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .webp | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .heic | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
| .avif | Phase B converter — OCR (design only): text is read from the image, translated, and the image is rebuilt — Live on the platform (extract-only) |
Pages, iWork & web archives (4)
| Format | Status / route |
|---|---|
| .pages | Phase B converter — iWork conversion (design only) — Live on the platform (extract-only) |
| .key | Phase B converter — iWork conversion (design only) — Live on the platform (round-trip) |
| .numbers | Phase B converter — iWork conversion (design only) — Live on the platform (round-trip) |
| .mhtml | Phase B converter — document converter (design only) — Live on the platform (round-trip) |
Counts: 104 formats in total — 103 live, 1 not supported (missing piece named on each row). A format moves to “Live” only when its import and export are proven on the project path on real files. No format is ever silently downgraded to plain-text garbage: an unsupported file returns its route, and a corrupt supported file returns its reason.
How this scales to 280+ pairs — by design, not by exception
- Templated per-pair configuration: each pair/dialect gets the same template (roster, TM, glossary, domain, QA profile, Human Gate) — no hand-built pair files, no pair-specific shortcuts.
- Per-pair and per-client TM/glossary: matches, leverage and payable shares are computed from that pair's memory, never borrowed from another pair.
- Same quality bar for every pair: 110 running QA checks in this browser (104 QA checks on the live platform), reviewer ≠ translator, and delivery blocked by a critical issue, an untranslated segment, or a missing Human Gate approval — for pair 1 and pair 280 alike.
- Same user experience for every pair: segment search and filters, TM-band filter, reference-field search, concordance, comments, version history, locking, tags as chips, and the keyboard shortcuts listed below work identically whatever the pair.
- Roster-before-sales, always: a pair with no verified translator + separate reviser shows WAIT — NOT SOLD and cannot be ordered, however large the language list grows. Growth in pairs never weakens the gate.
Today's confirmed base is 150+ languages, recruited and verified pair by pair, across 2,852 roster-approved pairs, and this page's sold roster is the 3 pairs named above.
The live Lingua CAT Tool in the platform has a backend, real engines, live file upload for 103 formats, and accounts. Your work here is auto-saved in this browser only (localStorage). Paid human-reviewed delivery is only sold when a verified translator and a different verified reviser exist for the pair (roster-before-sales) — otherwise the pair is WAIT — NOT SOLD.
How to use this editor
- Use Document Intelligence below the pair engine for a real document: drop the file in, let it detect the source language and document type, run its 10-point check, then apply the detected source to the pair engine. The target always comes from your requirement, never from a guess at the source text.
- Pick a pair in the Universal pair engine above and check its readiness — then Click a segment, edit its target on the right, and confirm it with Ctrl+Enter — or the Confirm button.
- On a live project, the Team panel above the segments shows who is online and which segment each person is on. Click a segment to claim it — a segment locked by someone else cannot be edited until they release it.
- Use the TM match shown under each segment: click Insert TM to paste the best match, then edit it.
- Click a gold glossary term in the source to insert its approved translation at your cursor.
- As you type a target, the editor suggests the next words from the translation memory, the glossary and the MT draft — Tab or Enter accepts a suggestion, Esc dismisses it; accepted suggestions are learned, and confirming a segment makes its pair a suggestion source for the next segment. The visual preview below renders the whole document — click any line to open that segment.
- Run QA any time; click a finding to jump straight to that segment and fix it.
- Delivery stays blocked while a critical QA issue or an untranslated segment is open — approve the Human Gate, then use Delivery.
- Split a segment at a sentence boundary or at your target cursor, merge it with the next segment, or join it with the previous one — the buttons sit on each segment. A split or merge clears the Human Gate approval and re-runs QA, and can be undone from the toolbar.
- Tags ({project_name}, email addresses, serial numbers) appear as chips in the source and QA flags any that are missing or changed in the target; use Transfer tags on a segment to insert any missing tag at your cursor, so a confirm is never refused for a tag you forgot to copy.
- Filter QA findings by content type (General, Software strings, Subtitles, Sensitive) and suggest new terms as PENDING — a suggestion is never auto-approved.
- Filter segments by TM band (New, fuzzy bands, Exact, Repetition) and search by source, target or reference field; each segment header shows its TM band and its payable share under the approved scheme.
Shortcuts: Ctrl+Enter confirm segment · Alt+Down / Alt+Up next / previous segment · Alt+1 insert best TM match · Alt+2 insert first glossary term. Work without the mouse: Tab moves into a target box, and the shortcuts move you between segments — but while a predictive suggestion is showing, Tab accepts it instead (Enter accepts it too, Esc dismisses it). · Alt+S split segment · Alt+M merge with next · Alt+J join with previous · Ctrl+Z undo split / join (outside target boxes)
Project dashboard — project, counts
File name: -project-01.txt · Pair: English (en) → Draft target (not a delivered language pair) · Domain: General / platform description · Currency of counts: segments in this browser only.
—
—
—
—
—
Progress by status
TM leverage breakdown — approved TM net-rate scheme, translation stage
Rates below are the approved scheme for the translation stage. Exact/repetition is 20% at translation stage and 100% at review stage; review-stage percentages for the other bands are not specified in the approved scheme and are shown as “—”, not invented.
| TM band | Segments | Source words | Translation stage | Review stage |
|---|
How bands are counted here (example only): a segment's band comes from its best TM match — 100% exact, 95–99%, 85–94%, 75–84%, 50–74% fuzzy, below 50% new. A segment whose source already appeared earlier counts as repetition, regardless of its TM score. Counts change as you edit segments, because TM scores are recomputed from the TM.
Segment workflow
Untranslated → Draft → Translated → Reviewed → Confirmed
Locked is a separate hold state, not a step in the flow: a locked segment cannot be edited until it is unlocked. Use the Advance status button on a segment to move it one step forward, or the status dropdown to jump to a step. Filter by status and search source or target in the project card above; comments and version history sit under each segment. Each segment also shows a reference field (a software string key, a subtitle timecode, or screenshot tag text) and tags appear as chips that QA checks for changes. Split, merge and join are on each segment; a split keeps the first half's number and gives the second a new number, and a merge keeps the first segment's number — numbers are never reused.
Keyboard shortcuts — only shortcuts this page actually supports
These shortcuts work when a segment target box is focused. No other shortcut is claimed on this page.
- Ctrl+Enter — confirm the focused segment (moves it to Confirmed and records a version in its history).
- Alt+Down — move to the next segment. Alt+Up — move to the previous segment.
- Alt+1 — insert the best TM match into the focused segment.
- Alt+2 — insert the first approved glossary term found in the focused segment's source.
- Tab — accept the predictive suggestion when one is showing; otherwise standard browser focus movement into and between fields.
- Enter — accept the predictive suggestion when one is showing; otherwise a newline in the target box.
- Esc — dismiss the predictive suggestion.
- Alt+S — split the focused segment (at a sentence boundary or at the target cursor).
- Alt+M — merge the focused segment with the next segment.
- Alt+J — join the focused segment with the previous segment.
- Ctrl+Z — undo the last split, merge or join (only when focus is outside a target box).
On some keyboards and browsers, Alt combinations may be intercepted by the browser or operating system; the buttons on each segment do exactly the same actions, so no feature is shortcut-only.
Project — document: "How LinguaEmpire works"
Source: English · Target: translations below. Edit them freely; every change is saved in this browser only.
·
“Changed” means the segment's source differs from its recorded previous version, or the segment was edited in this browser session. Segments with no prior version recorded are marked as such and appear under neither “Changed only” nor “Unchanged only”. This filter changes only what is shown: QA, the Human Gate and all counts still run on the full set.
Segment in context — active segment preview (RTL toggle)
Shows the active segment's target inside a simple layout, so you can see it in context before confirming.
Direction: left-to-right
Page
Select a segment to preview its target here.
String context — active segment
Context travels with the string: the card below shows only what the segment's own data carries — its string key, source file, and any developer note the parsed data supplied. Where none exists, it says so; no screenshot, UI location, or developer note is invented. Display only: it changes no segment, QA result, or gate decision.
Select a segment to see its context here.
Team & live collaboration — who is here, what they are doing
Work on a live project with your team: see who is online, which segment each person is on, comment on segments, ask questions, and follow assignments and workflow steps. This panel is live only when the editor is opened from a live project with a sign-in — on this demo page it shows why it is not connected, and the comments and locking below keep working in this browser only.
Demo — not connected
Online now
No one is shown as online — open a live project to see your team here.
Team & assignments
Assignments are shown when a live project is open.
Your token is kept in this tab's session storage only, sent only to this site's /api/v1, and never written into the page. Close the tab and it is gone.
Open questions
No open questions — ask one from any segment.
Project discussion
Project discussion appears when a live project is open.
Activity
No activity yet in this session.
Workflow
Workflow steps are shown when a live project is open.
Tracked changes — every edit, with accept / reject
Every edit you make to a target is recorded as a tracked change: deleted words are struck through, added words are highlighted. Accept keeps the edit; Reject restores the text before the edit. This extends the per-segment version history below each segment — history records confirmed snapshots, tracked changes record the edits between them. On this demo page changes live in this browser only; in a live project they are also stored on the server and shared with the team.
No tracked changes yet — edit a target to record one.
No tracked changes yet.
Version history + Word round-trip — live platform
Every edit in the live platform is a version step in the server version history: the full matrix (every step with its source), a red/green diff between steps, a screenshot of each step, and restore from any step. Word review works end to end: export a .docx with real tracked changes (added words green, removed words red), review it in Word, re-import it — the reviewed targets are applied as new version steps. This panel calls the live API (same token as the MT toolbar); it needs a platform project/segment UUID — the demo segments above are numbered 1–10 in this browser and have no UUID.
No version history loaded yet.
Word export/import runs against the live platform. A file whose source text does not match the project is refused, never silently re-paired.
LiveDocs — reference material in the editor
Your LiveDocs reference documents, inside the CAT editor: list a client's documents, open one, and search its concordance (every sentence containing a phrase, with its aligned translation). This panel calls the live platform API (same token as the version-history panel above); it needs platform UUIDs — the demo segments above are numbered in this browser and have no UUID.
No LiveDocs listed yet.
List first, open a document, then search its concordance.
Side-by-side review — second editor view
Every loaded segment as a source | target row with checkboxes, for review in bulk — Side-by-side editor mode. Tick segments, then approve them all at once or copy their sources into the targets. This is a view of the segments already loaded in this browser; the bulk actions use the same per-segment actions as the buttons above.
Side-by-side view is off — it shows every loaded segment as source | target rows with checkboxes, for review in bulk.
Concordance search — TM only
Select or type a word or phrase to see every TM unit that contains it.
Shared TM in this session: when you confirm a segment, it joins the in-session TM, so the same segment worked on in this browser session is found by concordance search afterwards. Sharing runs in this browser session.
Glossary / termbase manager — terms, localStorage only
Approved terms appear in gold in the source. Forbidden terms are flagged in QA. Do-not-translate (dnt) terms must stay exactly as written. Edit loads a term into the fields above; Delete removes it from this browser only. “dnt” is shown in full as “do-not-translate” in the list.
Suggest a term — stays PENDING until a reviewer approves it
A suggested term is never auto-approved: it is not highlighted in the source and not enforced in QA until a reviewer approves it. Termbase scope on this page is this project only, in this browser; a shared, per-client termbase runs across projects in the live platform.
Translation memory & glossary sources — working locally now, online in Phase B
What works now, in this browser: import a TMX or TBX file you supply. Entries are counted instantly here and kept in this browser only (localStorage) — they are never sent anywhere, and they never replace the TM and glossary on the segments below. What does not work from this static page: any live online TM or glossary connection. A static page has no server-side key, no account and no licence to call an external service with, so any claim of a live online connection here would be false. Real-time in Phase B means the hosted backend serves the pair's own TM/glossary on every save; here, “instant” means instantly inside this browser session.
Imported TM units: 0 (browser-local only)
Imported glossary terms: 0 (browser-local only)
Only files you choose are read. Imported entries are used by checks 8 and 9 in the Document Intelligence panel.
| Source | Status on this page | What Phase B needs before it goes live |
|---|---|---|
| LinguaEmpire per-client TM & glossary | Example only (segments below) | Hosted backend + accounts + per-client databases; proven on real delivered work |
| TMX / TBX files you supply | Browser import live now (counts above) | Server-side import with validation, deduplication and versioning |
| Public terminology services (for example IATE, UNTERM) | NOT CONNECTED — no data copied | Backend proxy, each service's terms/licence check, rate limits, attribution |
| Licensed provider glossaries & MT memories | NOT CONNECTED | Licensed API key held server-side only (never in a browser page), contract, cost approval |
| Community / open TM collections | NOT CONNECTED | Licence review per collection — open does not mean free to resell or to claim as LinguaEmpire data |
Nothing in this table is a live connection today. A source moves from “NOT CONNECTED” to live only when its backend, key, licence and cost approval all exist and are verified — not when a row is added to a table.
Import — paste text or a .txt file (read in this browser only)
Imported text replaces the grid for this session only: no original formatting is preserved, and no file is sent anywhere. TMX/TBX import lives in the Sources panel below (browser-only, from files you supply); XLIFF import runs in the live platform.
Visual context preview — see the translation inside a page, not just in a grid
Translators make better choices when they see where words live. This preview renders the whole document — every segment in order, with the active segment outlined; click any line to open that segment. Press Refresh after confirming a segment to see its target in place.
It shows where the translated words sit. It is not a published website and it does not publish anything.
Rendered-document preview — InDesign (.idml)
Load an .idml file to see the translation inside its original InDesign pages: each page is drawn with its text frames in place. Frames with a matching segment translation show the translation; the rest show the source text with a dashed outline. The file is read in this browser when the browser can unzip it, otherwise through the platform preview (POST /api/v1/cat/idml/preview) when a sign-in token is present — nothing is sent anywhere without the platform route, and the file itself is never stored.
Document view — all segments in order.
Productivity, TM leverage & quality estimates — example data, honestly labelled
Counts below are computed live from the current segments. The words-per-hour figure is a session estimate from this browser session only — it is not a productivity claim about any translator, and it is never used to set pay.
Per-segment quality estimates
Press Refresh statistics to see estimates.
An estimate blends the TM match score with glossary coverage. It is a planning aid, not a quality guarantee: the Human Gate and the QA checks (110 in this browser; 104 on the live platform) decide delivery, never the estimate.
Workspace customisation — font, density and theme
Every translator works differently, so the workspace adapts without changing any translation data. These controls work in this browser only and are remembered here (localStorage).
Customisation changes presentation only — segments, TM, glossary, QA and the Human Gate are untouched.
Panels — show or hide (modular workspace):
Hiding a panel hides its display only — its data, QA findings, glossary terms and the Human Gate are untouched, and every feature stays available from its own button or card when shown again. Panel choices are remembered in this browser only.
Integrations & AI copilot
These connections run in the live platform, each with its backend, key, licence and cost approval.
| Connection | Status | What it needs before it goes live |
|---|---|---|
| Code hosting sync (GitHub, GitLab) for continuous software localisation | NOT CONNECTED | Hosted backend, repository token held server-side, webhooks, branch rules |
| Design tool sync (Figma) for in-context strings | NOT CONNECTED | Licensed API access, file/frame mapping, designer approval flow |
| Content systems (CMS) for web content | NOT CONNECTED | Per-system connector, content-type mapping, publishing approval (human-only) |
| Machine translation providers | NOT CONNECTED from this page | Licensed API keys server-side only, per-pair engine choice, post-editing workflow (a machine draft is never a delivery) |
| AI copilot drafts | DESIGN ONLY | Drafts built from the pair's TM, glossary and approved style; every draft needs human confirmation; copilot never confirms, delivers or signs anything |
| Quality estimation | (estimates card above, example data) | Calibration on real jobs and measured LQA before any estimate is shown to a client |
Publishing, spend, account access and claims stay human-only, as they do everywhere else in this platform.
QA dashboard — counts, per-segment panel, and QA before delivery
Run QA first. Counts below are for the current segments only.
—
—
—
—
Per-segment QA panel
Run QA to see per-segment findings.
QA before delivery — checklist (delivery stays blocked until every required item passes)
- Critical QA issues cleared: not yet checked — run QA.
- No untranslated or empty segments: not yet checked.
- Human Gate decision recorded and approved: no.
- LQA score: not yet computed — run QA. Threshold is 90.
- QA run completed (all check lanes settled): no — run QA.
“Cleared” means the issue no longer appears when QA is re-run after fixing the segment — it is not a button that dismisses a finding. A gate approval never overrides a critical issue or an untranslated segment on this page.
QA findings — every finding names its rule and segment
QA by content type (presets — filtering only):
Unchecking a type hides its findings from this list only. It does not clear them, and delivery still counts every critical finding.
Run QA to see findings.
AppStore — CAT plugins
The plugin store for this editor: installed plugins with their manifests, permissions and extension points, and enable / disable. Plugins run in-process behind an explicit permission manifest — no remote code is fetched and no plugin code is uploaded; registration admits only in-tree modules. Plugin QA findings merge into the QA dashboard below exactly like any other finding, and a plugin that fails is reported as a plugin-error finding, never a silent pass.
Not loaded yet — sign in (Team panel token) and load the store; nothing is listed without a live answer.
Not connected. The installed plugins, their manifests and their enablement state come from the live platform only.
Try a plugin on the active segment
No plugin action tried yet. The segment action proposes text — Apply sends it through the normal edit path (tracked changes + version history), the plugin itself never writes the segment.
QA checks — the full list (110 checks run in this browser)
Every check below runs in this browser editor when you press Run QA check. Severity critical blocks delivery at the Human Gate; warning does not. Checks marked with a language in their name only apply to that language; all others apply to every pair. The Spelling check calls the live QA service (POST /api/v1/qa/spellcheck); if the service is unreachable, or the language has no dictionary, the segment is reported as “Spelling not verified” — never as a pass.
Honest count: this editor runs 110 checks. The live platform’s backend QA catalogue holds 104 checks, all wired into this browser editor (0 unwired). The full backend catalogue is at GET /api/v1/qa/catalogue. Your own custom regex checks, if any are defined in the Custom QA panel above, run in addition to the 110 listed here.
| Check | Category | Severity |
|---|---|---|
| Untranslated segment / Empty target | Accuracy & Completeness | critical |
| Missing final punctuation | Linguistic Conventions & Fluency | warning |
| Numbers mismatch | Technical & Formatting | critical |
| Tag mismatch | Technical & Formatting | critical |
| Casing | Technical & Formatting | warning |
| Terminology deviation | Terminology & Consistency | critical |
| Forbidden term | Terminology & Consistency | critical |
| Do-not-translate term | Terminology & Consistency | critical |
| Consistency | Terminology & Consistency | warning |
| Length | Accuracy & Completeness | warning |
| Leading / trailing spaces | Technical & Formatting | warning |
| Double spaces | Technical & Formatting | warning |
| Repeated word | Linguistic Conventions & Fluency | warning |
| Bracket mismatch | Technical & Formatting | critical |
| Space before punctuation | Linguistic Conventions & Fluency | warning |
| Sentence count differs | Linguistic Conventions & Fluency | warning |
| Target identical to source | Accuracy & Completeness | warning |
| Date format differs | Linguistic Conventions & Fluency | warning |
| URL mismatch | Technical & Formatting | critical |
| Format specifier mismatch | Technical & Formatting | critical |
| Percent sign mismatch | Technical & Formatting | warning |
| Acronym / all-caps term missing | Technical & Formatting | warning |
| Tag text left in target | Technical & Formatting | warning |
| Imported glossary term not applied | Terminology & Consistency | warning |
| Spelling | Linguistic Conventions & Fluency | critical |
| Grammar — lowercase pronoun "I" | Linguistic Conventions & Fluency | warning |
| Grammar — missing apostrophe in contraction | Linguistic Conventions & Fluency | warning |
| Grammar — wrong ordinal suffix | Linguistic Conventions & Fluency | warning |
| Grammar — sentence starts lowercase | Linguistic Conventions & Fluency | warning |
| Grammar — sentence ends mid-phrase | Linguistic Conventions & Fluency | warning |
| Grammar — missing apostrophe (French elision) | Linguistic Conventions & Fluency | warning |
| Grammar — "a"/"an" mismatch | Linguistic Conventions & Fluency | warning |
| Quote mismatch | Technical & Formatting | critical |
| Repeated punctuation | Linguistic Conventions & Fluency | warning |
| Email mismatch | Technical & Formatting | critical |
| Malformed URL | Technical & Formatting | warning |
| Broken HTML entity | Technical & Formatting | critical |
| Missing space after punctuation | Technical & Formatting | warning |
| Invisible character | Technical & Formatting | warning |
| Double capital at word start | Technical & Formatting | warning |
| Full-width character | Technical & Formatting | warning |
| Bracket count differs | Technical & Formatting | critical |
| Missing opening ¿ / ¡ | Linguistic Conventions & Fluency | warning |
| Time missing from target | Linguistic Conventions & Fluency | critical |
| Measurement unit missing | Linguistic Conventions & Fluency | critical |
| Phone number mismatch | Technical & Formatting | critical |
| Version number mismatch | Technical & Formatting | critical |
| Mixed apostrophe style | Linguistic Conventions & Fluency | warning |
| Ellipsis missing from target | Linguistic Conventions & Fluency | warning |
| Space before % | Linguistic Conventions & Fluency | warning |
| Missing space before bracket | Technical & Formatting | warning |
| Space inside brackets | Technical & Formatting | warning |
| Double hyphen used as dash | Linguistic Conventions & Fluency | warning |
| Space inside quotes | Technical & Formatting | warning |
| French guillemet spacing | Linguistic Conventions & Fluency | warning |
| English month/weekday untranslated | Accuracy & Completeness | critical |
| File name missing or changed | Technical & Formatting | critical |
| Line breaks differ | Technical & Formatting | warning |
| Parenthetical left untranslated | Accuracy & Completeness | critical |
| Mixed decimal separators | Linguistic Conventions & Fluency | critical |
| Unbalanced single quote | Technical & Formatting | critical |
| Repeated sentence | Linguistic Conventions & Fluency | warning |
| Mixed date separators | Linguistic Conventions & Fluency | warning |
| Currency code glued to number | Technical & Formatting | warning |
| Redundant currency (symbol + code) | Linguistic Conventions & Fluency | warning |
| English number word untranslated | Accuracy & Completeness | critical |
| Wrong decimal separator for locale | Linguistic Conventions & Fluency | critical |
| Currency marking missing | Linguistic Conventions & Fluency | critical |
| Missing space before French punctuation | Linguistic Conventions & Fluency | warning |
| Missing space before % (French-style) | Linguistic Conventions & Fluency | warning |
| Mixed curly/straight quotes | Technical & Formatting | warning |
| Target all in capitals | Technical & Formatting | warning |
| Target has no capital letter | Linguistic Conventions & Fluency | warning |
| Month/weekday lowercase in English | Linguistic Conventions & Fluency | warning |
| Ordinal missing from target | Linguistic Conventions & Fluency | critical |
| Em dash spacing (French) | Linguistic Conventions & Fluency | warning |
| Number range reversed | Technical & Formatting | critical |
| Word capitalised inconsistently | Technical & Formatting | warning |
| Unit capitalisation differs | Technical & Formatting | warning |
| Weekday/month swapped | Linguistic Conventions & Fluency | critical |
| English abbreviation untranslated | Linguistic Conventions & Fluency | warning |
| Currency precision differs | Technical & Formatting | critical |
| Missing space between number and unit | Technical & Formatting | warning |
| Currency symbol on wrong side | Linguistic Conventions & Fluency | warning |
| AM/PM swapped | Linguistic Conventions & Fluency | critical |
| Minus sign dropped or added | Technical & Formatting | critical |
| Roman numeral differs | Technical & Formatting | critical |
| Lookalike character from another script | Technical & Formatting | critical |
| Placeholders in different order | Technical & Formatting | critical |
| Term translated differently elsewhere | Terminology & Consistency | critical |
| Style guide violation | Terminology & Consistency | warning |
| Negation mismatch (smell) | Accuracy & Completeness | warning |
| Unmodified fuzzy TM match | Accuracy & Completeness | critical |
| Target over character limit | Accuracy & Completeness | critical |
| Target length outside profile bounds | Accuracy & Completeness | warning |
| Copied from source | Accuracy & Completeness | critical |
| Unedited machine translation | Accuracy & Completeness | critical |
| Forbidden character | Technical & Formatting | critical |
| Forbidden pattern | Technical & Formatting | critical |
| Required pattern missing | Technical & Formatting | critical |
| Wrong script for target language | Linguistic Conventions & Fluency | critical |
| Length ratio outside pair band | Accuracy & Completeness | warning |
| Placeholder missing or changed | Technical & Formatting | critical |
| First word not capitalised | Linguistic Conventions & Fluency | warning |
| Brand name missing or changed | Terminology & Consistency | critical |
| Semantic error (AI) | Accuracy & Completeness | critical |
| Grammar (AI) | Linguistic Conventions & Fluency | warning |
| Omission (AI) | Accuracy & Completeness | critical |
| Addition (AI) | Accuracy & Completeness | critical |
| Tone / register (AI) | Style & Readability | warning |
All 104 backend catalogue checks are wired into this browser editor (0 unwired). Full catalogue at GET /api/v1/qa/catalogue.
Human Gate — delivery block
Approver: the pair's roster linguist — the named native linguist for that pair — approves at the Human Gate. LQA threshold: 90. The Deliver button stays disabled while any critical QA issue is open or any segment is untranslated, even after a gate approval, and it also requires a recorded gate approval. A “Return for rework” decision is a recorded decision, but it does not enable delivery. Delivery also requires a completed QA run.
Export preview
JSON and TMX-style downloads are generated in this browser from the segments only and are labelled “ EXPORT” inside the file. They are not real client files, not delivered work, and no file is sent anywhere.
How this compares, honestly
The four named tools below are live products. LinguaEmpire's editor on this page is a working version — it proves the interaction in the browser. The live Lingua CAT Tool in the platform has a backend, a real TM database, real MT engines, a roster of 2,852 pairs and real delivered jobs; this walkthrough shows the interaction only.
| Feature category | Leading live CAT products | LinguaEmpire on this page |
|---|---|---|
| Segment grid | Live | Working version — 10 segments, editable, statuses, progress |
| Translation memory | Live, shared databases | Working version — TM, exact/near matches, concordance, TMX-style preview |
| Glossary / termbase | Live, enforced | Working version — gold highlights, click-to-insert, forbidden and do-not-translate flags |
| Automated QA | Live, whole-file | Working version — 110 checks, severity, filter, jump-to-segment, content-type filter |
| Visual context | Live, in-context review | Working version — rendered-document preview (all segments in flow, click to open a segment; IDML pages when a file is loaded) with RTL toggle |
| Keyboard shortcuts | Live | Working version — confirm, move, insert TM, insert glossary term |
| Collaboration | Live, multi-user | Live, multi-user in the platform — presence, live cursors, server comments, queries, assignment, workflow steps; roles, comments and locking in this browser on this demo page. |
| File formats | Live, many formats | — pasted text/.txt import, Document Intelligence reads .txt/.md/.csv/.tsv/.srt/.vtt/.xml/.xlf/.json/.html in-browser, TMX/TBX import in-browser from files you supply (browser database, no 500 cap); DOCX/XLSX/PPTX/ODT/ODS/EPUB/RTF parsed in-browser by the built-in reader; PDF is live in the platform (text-layer import + new-layout export; scanned-PDF OCR is not built); all 104 target formats are listed with honest routes in the Formats & converters panel; online TM/glossary connection in the live platform |
| Tags | Live, protected in most editors | Working version — variables, serial numbers and email addresses shown as chips, QA flags missing or changed tags |
| Reference fields | Live, per segment | Working version — string key, subtitle timecode, or screenshot tag text shown on every segment |
| Term suggestions | Live, with approval workflow | Working version — suggestions stay PENDING, never auto-approved, until a reviewer approves them |
| Shared TM | Live, team-wide | A confirmed segment joins the in-session TM in this browser. |
| Machine translation | Live, multi-engine | Working version — live MT suggestions from every engine this server reports as configured (no fixed list), side-by-side compare of up to 3 engines with a quality estimate on each draft, glossary-aware MT from the approved termbase, and the engine shown on every MT draft. A machine draft is never a delivery: it still needs human editing, QA and confirmation. |
| Document alignment | Live, aligners in leading tools | Working version — Document Aligner: sentence alignment (length-based + anchor matching), editable pairs, TMX export, add to CAT TM |
| Tracked changes | Live, with changes pane and accept / reject | Working version — per-segment tracked changes with insertions/deletions, accept/reject, changes pane by workflow step, HTML export |
| Segment split / join / merge | Live | Working version — split at a sentence boundary or target cursor, merge with next, join with previous, undo, QA and TM re-run on the affected segments |
Where the live CAT Tool goes further: the Human Gate delivery block is live (delivery cannot be forced past a critical QA issue), per-segment explainable QA with reasons runs on the 104 live QA checks, and per-client TM and glossary run in the live editor — proven on real jobs.
Frequently asked questions — in LinguaEmpire's own words
What is a CAT tool, and how is it different from machine translation?
A CAT tool is an editor for a human translator. It splits a document into segments, reuses past translations from a translation memory, enforces an approved glossary, and runs quality checks — but the human writes, edits and confirms the translation. Machine translation alone is a machine draft with no memory, no glossary enforcement and no human confirmation. This page is a CAT editor : its drafts are labelled examples, not machine deliveries.
How does the Lingua CAT Tool fit into the LinguaEmpire platform?
The Lingua CAT Tool is the editor inside the LinguaEmpire platform. Around it, the platform manages projects, linguists, deadlines, invoicing and delivery — the editor is where the translation work happens, and the platform is the system that runs the whole job from order to delivery.
Which file formats can I import here?
On this page, pasted text and .txt files work in the Import panel, and the Document Intelligence panel reads .txt, .md, .csv, .tsv, .srt, .vtt, .xml, .xlf/.xliff, .json and .html in this browser, detects the source language, classifies the document type, and runs a 10-point check. TMX/TBX files you supply yourself can be imported in the Sources panel in this browser only. .docx, .xlsx, .pptx, .odt, .ods, .epub and .rtf (basic reader) are parsed in-browser by the built-in reader; .pdf is live in the platform (text-layer import + new-layout export) but is not read by this browser panel. The Formats & converters panel lists all 104 target formats with their honest routes, including OCR routes for images and scanned PDFs. No external TM or glossary is connected live from this page.
Do I need to install anything?
This page runs in the browser on any operating system and stores work in this browser's localStorage only. If you clear this browser's stored data, your work on this page goes with it.
What happens when I confirm a segment?
It moves to Confirmed, a version is recorded in its history, and it joins the in-session translation memory, so concordance search in this browser finds it afterwards. That sharing runs in this browser session — presence is built into the live editor.
Who approves a suggested term?
A reviewer, using the Approve button next to a pending suggestion. A suggestion is never auto-approved: until then it is not highlighted in the source and not enforced in QA.
What blocks delivery?
Three things together: a recorded Human Gate approval (approver: the pair's roster linguist, the named native linguist for that pair; LQA threshold 90), zero critical QA issues, and zero untranslated segments. A gate approval never overrides a critical issue or an untranslated segment.
Scale & large files — what this editor handles today, and what Phase B adds
Built today (this page): the segment editor pages its display — 50 segments per page — so a large document no longer floods the browser with thousands of segment cards at once. Filters, search, QA (110 checks), the Human Gate and all counts still run on the full segment set, not just the visible page. Document Intelligence analyses pasted or file text in this browser; text over 1,000,000 characters is analysed on a labelled (first 400,000 + last 100,000 characters) and the is stated on the result, never hidden.
Honest caps that remain: this page has no server, so it cannot run a full PDF engine in-browser, and editor segment sets still live in browser memory with a best-effort local save — a very large set may not persist across reloads. Imported TM and glossary are NO LONGER capped at 500: they now persist in the browser's own database (IndexedDB) with local storage kept only as a 500-entry cache for older browsers. The built-in example data stays and labelled — but the working set is no longer -only: your DOCX/XLSX/PPTX/ODT/ODS/EPUB/VTT/TSV/RTF documents, your TMX/TBX imports and your glossary are now workable editor data in this browser. Live/shared data in the platform: accounts, verified roster, real jobs, payments and online connections. A file that needs a live engine, a live account or a licence runs through the live platform.
Platform design: chunked, resumable uploads with a checksum per chunk; streaming parsers that read a file piece by piece instead of loading it whole; background job queues with retries so a long job never blocks the editor; server-side pagination and virtualised rows in the editor; indexed TM and glossary search (per pair, per client) instead of scanning lists; object storage for files and a database that scales by client and pair. That design is how the platform reaches no practical limit for real translation workloads — a literal limit-free system does not exist, and the design states its ceilings (chunk size, job timeouts, per-plan fair-use quotas to be confirmed) instead of hiding them. The same design covers the Platform Core project editor, whose table today is -only.
Where this page falls short — stated honestly
This page is a working browser walkthrough, not a finished product. Its honest limits are:
- The 110-check QA engine is live in the editor. QA runs 110 checks, including numbers, tags, terminology, consistency, punctuation, casing, spacing, length, empty targets, repeated words, bracket balance, spacing before punctuation, sentence count, dates, URLs, format specifiers, percent signs, acronyms, leftover tag text, imported-glossary terms, spelling (live Hunspell service), grammar, and the full list published below.
- Full offline work is not available. The page needs a browser and stores work in localStorage only; it does not work as a standalone offline desktop tool.
- Presence is built in the editor. Editor presence (who is viewing which segment) is built into the live CAT editor and shown there. Locking, comments, roles and version history work in this browser.
- Trial credits and billing run in the live platform. The live platform has real machine-translation engines, a real translation-memory database, a roster of 2,852 pairs and ordering. Trial credits and billing are live in the platform; card payment receipt is not connected yet. Start a job in the live platform.
Trial credits and billing are live in the platform. The 1 unsupported format is listed with the missing piece named on each row. The live platform already has the backend, per-client memory and glossary, and real file import for 103 formats — each proven on real data.
Original LinguaEmpire content; no competitor text, code, logo or design is copied on this page.