Control
Pick the model that fits the work
Different models read different languages differently. Claude for nuance and long context. Gemini for European pairs. GPT for marketing copy. Groq for fast first drafts. Transept gives you all of them — pick per document, switch per sentence, or let Premium and Publish-Ready modes route automatically.
Free to begin · No card required
Different AI models read different languages differently. Claude tends to handle long-context nuance and literary register particularly well; Gemini is strong on European technical pairs; GPT excels at marketing copy and persuasive tone; Groq runs the fastest cheap drafts when you’re moving quickly. Locking into one provider means accepting that one model’s strengths and weaknesses on every language pair, every register, every document type. Transept routes through all of them — Claude, Gemini, GPT, Groq, and OpenRouter for everything else — and lets you pick the right model per document, per sentence, or by mode.
One workspace, every frontier model
Single-provider tools lock you into one model’s quirks. Transept routes through Claude, Gemini, GPT, Groq, and OpenRouter — and you decide which one runs.
Pick per document
Set a default model when you open a new document. Claude Opus for literary; Gemini for German technical docs; GPT for marketing.
Switch per sentence
A sentence struggling under one model often comes out clean under another. Regenerate with a different model from the same alternatives panel.
Save presets
Model + temperature + prompt template, saved as a preset you can apply with one click on the next document.
Premium routes automatically
Premium and Publish-Ready modes pick the strongest available model for the language pair. Standard mode runs the fastest reasonable choice.

Why this is worth caring about
No vendor lock-in
When a better model ships, you get access. No re-platforming, no contract renegotiation.
Right tool for the language
Some models are stronger on certain pairs. Picking by pair beats picking by brand loyalty.
Cost control
Use the fast cheap model for first drafts; bring out the heavy one for the chapters that need it.
Questions, answered without the fluff
- Claude Opus, Claude Sonnet, Claude Haiku; Gemini Pro and Flash; GPT-4 and -4o; Groq Llama; plus OpenRouter for everything else. The list grows when new frontier models land.
- No — every model is included in the credit price for the mode you pick (Standard 1×, Premium 3×, Publish-Ready 5×). The model choice is yours; the credit cost is the same.
- Only the one you’ve picked. The other adapters are inactive for that translation. See the security page for the full data story.
- Regenerate the sentence with each model and the alternatives panel shows both. Pick the winner.
- Depends on the pair and the register. For literary EN→FR, Claude Opus tends to win. For German technical content, Gemini Pro often does. For marketing copy in Romance languages, GPT-4 family. The honest answer: try a paragraph under each, compare in the alternatives panel, pick.
- No. Transept handles the provider relationships and bundles model access into your credit price. You don’t manage keys, rate limits, or provider accounts.
- OpenRouter is the gateway for everything beyond the four direct providers. If a model is available on OpenRouter, it’s accessible. Premium and Publish-Ready modes route to the best available model automatically; for narrow control, set the model explicitly.
- Only the one you’ve picked for that translation. Other adapters are inactive for that request. See the security page for the full data-handling story per provider.
- Save a combination (model + temperature + prompt tweaks) as a preset; apply with one click on the next document. Most translators have 2–4 presets — one per client, one per content type — that they reuse for months.
Keep reading around this capability
- FeatureSentence-level alternatives→Translation isn’t one right answer. Every sentence has three or five reasonable choices, and the best one depends on the page around it.…
- FeaturePublish-Ready Mode in one run→The whole workflow, one button. Translate. Enforce the glossary. Apply the styleguide. Proofread for drift. Polish for readability. Run QA.…
- FeatureA styleguide that steers every regeneration→Tone, register, rules, formatting conventions. Build a styleguide manually or have Transept generate one from a sample page of finished…
- CompareTransept vs ChatGPT for translation→ChatGPT, Claude, and Gemini are remarkable general-purpose AI assistants. They translate, summarize, rewrite, and explain — all from a chat…
- CompareTransept vs DeepL on the long documents→DeepL is the default for fast, high-quality AI translation. Transept doesn’t claim to beat DeepL on the raw translation of a single…