SSegmenta

AI translation platform

Fully automated translation workflow.Final human sign‑off.

Upload any document, apply your memories and termbases, and run AI drafting, revision and quality assurance — then review, refine and export in the exact original format.

  1. 01 Segment

    Files are split into segments with tags and placeholders protected.

  2. 02 Translate

    AI translation that follows your project terminology.

  3. 03 Revise

    Bilingual revision against the source meaning.

  4. 04 Quality check

    MQM-based LQA scoring plus automated tag, number and term checks.

Sign in

Your projects are private to your account.

or

No account?

Simple pricing

One workflow, sized for your volume

Every plan includes translation, AI revision and MQM quality checks. Start with 14 days free.

Starter

AI translation, revision and MQM quality check.

€49/ month
  • 50,000 words per month
  • 1 seat included
  • 2× words with your own engine keys

Pro

Most popular

Unlimited translation memories, per-project engine overrides, custom style guides.

€149/ month
  • 200,000 words per month
  • 1 seat included
  • 2× words with your own engine keys

Team

Shared memories and glossaries, admin cost auditing, 3 seats included.

€349/ month
  • 500,000 words per month
  • 3 seats included
  • 2× words with your own engine keys

No payment required to start. Cancel a paid plan anytime.

Why Segmenta

A controlled workflow, not a black box

Every stage is separate, visible and measurable — from segmentation to the file you download.

Document & tag fidelity

Native format in, native format out.

Translate complex layouts without losing formatting, styles or tag structures.

  • Every format you work in. Word, Excel, PowerPoint, XLIFF, HTML, JSON, Markdown and scanned PDFs (OCR).
  • Pre-translated CAT files. Import bilingual .xliff or .sdlxliff files with existing translations — pre-translated segments keep their targets and go straight to AI revision and QA.
  • Protected inline markup. Tags like <g1>…</g1> and <x1/>, variables and placeholders are isolated and guarded through every AI step.
  • Deterministic export. The original document structure and segment order come back exactly as they went in.

The three-stage pipeline

Three independent passes, zero black-box drift.

Separating drafting, revision and QA keeps accuracy and accountability at every step.

  • Draft translation. High-speed machine translation with automatic reuse of 100% memory matches — pre-translated segments bypass this stage automatically.
  • Bilingual revision. Surgical, minimal-edit revision against the source meaning — tone and syntax fixed, nothing rewritten for its own sake.
  • Automated & AI QA. MQM-categorized inspection (Critical, Major, Minor) checking numbers, terminology, tag integrity and cross-segment consistency.

Online human editing

A human workbench, right in the browser.

Review and edit every segment inline — AI stages and human decisions stay in one place.

  • Inline segment editing. Review any segment and edit the target directly in the project view before export.
  • Tag guardrails. Inline tags appear as compact chips and stay protected while you edit, with a warning before any removal.
  • Repetition syncing. One click adopts your preferred translation across identical repeated segments, keeping the whole document consistent.

Reusable assets

Your language assets stay private and improve every run.

Leverage past work without fine-tuning or leaking data to public models.

  • Industry formats. Upload .sdltm, .tmx and .xliff files directly into your account's private memory.
  • In-context guidance. Relevant past segments are dynamically retrieved and supplied to the AI as reference examples.
  • Glossary enforcement. CSV, TSV, TXT or XLSX termbases with inflection-tolerant checking keep vocabulary brand-consistent.
  • Client isolation. Assign specific memories per project so client assets never mix.

Engines & BYOK

Industry-leading MT and LLM engines, on your terms.

Pick the right engine per project or connect your own API accounts.

  • Engine choice. ModernMT, DeepL, Google Translate or AI drafting, combined with Claude and GPT models.
  • Account & project overrides. Set global account defaults, override engines per project, and attach a custom style guide that steers the revision pass.
  • Transparent BYOK. Use the included platform allowance or bring your own Anthropic and OpenAI keys to double your monthly quotas.

Quality metrics

Concrete data on every segment.

Know exactly what changed, how much was saved and where issues occurred.

  • Word-level diffs. Visual before-and-after diffs showing every edit made during revision.
  • Post-editing distance. Measured edit shift and post-editing distance metrics per document.
  • Human intervention rate. The exact share of segments accepted unchanged versus human-edited or verified — transparent proof of oversight.
  • Actionable QA reports. Drill into flagged issues, repetition conflicts and glossary compliance before exporting.

Start now

Try the full workflow free for 14 days

Translation, revision and quality checks on your own documents — no credit card required. Bring your own keys whenever you want.