CatScribe Beta Coming Soon
Windows and Linux beta builds are on the way.
Offline CAT translation workspace — built for books, subtitles, and narrative media.
Translate EPUB, DOCX, PDF, and subtitles with glossary-driven workflows, timing-aware tools, and local AI.
A CAT (Computer-Assisted Translation) tool helps translators work with glossaries, segments, and consistency — AI assists; you stay in control.
✅ 100% Offline-first
⚡ Large book workflow
📚 Glossary groups
🧠 Translation consistency
🎬 Subtitle synchronization
🧩 Dialogue-aware translation
🎨 Illustrated workflows (coming later)
Offline-first · Privacy-focused · Local AI powered
Features
Quick Translate
Fast snippet translations using local engines with clean UX.
File Translation
Chunked translation for EPUB/DOCX/PDF with progress + stability.
Glossaries & Memory
Keep names, places, spells consistent across an entire series.
Subtitle Translation
Translate SRT and VTT subtitles with professional tools: timing-aware chunk translation, CPS validation, dialogue continuity, auto-sync alignment (optional Whisper-based).
LM Studio Integration
Use LM Studio as a local provider to run compatible models with the same offline CAT workflow and glossary controls.
Illustrated Translation Coming Later
Future workflow for manga, comics, and illustrated storytelling: panel-aware translation, dialogue bubble workflows, visual narrative context.
Privacy & telemetry
Telemetry is anonymous and used only to improve performance and stability. You can opt out in-app at any time.
What is CatScribe Translate
CatScribe Translate is a CAT (Computer-Assisted Translation) tool: an offline AI translation workspace and desktop application for translating books and long documents with glossary consistency, translation memory, and per-segment editing so humans can refine the output.
Offline AI Translation Software
- Offline AI translation software — runs on your device; no cloud required.
- Translate books with AI using local models (e.g. Ollama) or optional cloud APIs.
- EPUB translation tool with DOCX and PDF support; chunk-based processing for long documents.
- Local AI translator with glossary consistency and translation memory across projects.
- CAT workflow: review and edit each segment — automatic translation isn’t perfect.
Who is CatScribe for
- Translators and authors working on books or long-form content.
- Translators working on long novels or fantasy series with recurring names and terminology.
- Technical documentation workflows requiring glossary consistency.
- Glossary-heavy projects (e.g. legal, medical, or brand terminology).
- Anyone preferring privacy-first, offline-capable translation instead of cloud-only tools.
Support CatScribe development ♥
If this tool helps you, consider supporting the project.
How CatScribe Works
Chunk-based processing
Split long docs into stable segments for reliable translation.
Local models
Run translation on your machine (Ollama, etc.).
Human-in-the-loop editing
Review and edit each segment to reach publication quality.
Glossary enforcement
Consistent terms across the entire book.
Translation memory
Reuse past translations for consistency.
Offline privacy
No data leaves your device.
Beyond Books: Narrative Media Translation
Subtitle workflows
Timing offset + stretch, dialogue consistency, episode-level memory, offline subtitle workflows.
Illustrated workflows
Manga workflows, comic translation, visual context awareness.
Why Offline Translation Matters
Privacy
Your documents never leave your device. No cloud uploads, no third-party access to your text.
Cost predictability
No per-token surprise bills. Run local models or choose when to use paid APIs.
Large-book workflows
No paste limits or context caps. Translate whole EPUBs, DOCX, and PDFs in one flow.
Consistency vs chat tools
Glossary and translation memory keep names and terms consistent — unlike ad-hoc chat prompts.
Why CatScribe is NOT a chat-based translator
CatScribe Translate is a CAT tool with file-in, glossary-driven workflows and segment-level editing. Chat-based tools (e.g. ChatGPT, DeepL paste) work one-off without glossary enforcement or translation memory.
- CatScribe: upload EPUB/DOCX/PDF → glossary + translation memory → segment review → consistent output.
- Chat/DeepL: paste text → one-off translation; no glossary, no segment editing, no book-length workflow.
- CatScribe runs offline with local AI; you stay in control of every segment.
Compare with other tools
Unlike DeepL or ChatGPT workflows, CatScribe Translate provides glossary consistency and book-length workflows offline.
| Product | Offline capable | Book-length workflows | Subtitle synchronization | Glossary consistency | Local models | Cost predictability |
|---|---|---|---|---|---|---|
| CatScribe Translate | Yes | Yes | Yes | Yes | Yes | Yes |
| DeepL | No | Partial | No | No | No | Partial |
| ChatGPT | No | No | No | No | No | Partial |
| LM Studio workflows | Yes | Partial | No | No | Yes | Yes |
App screenshots
These are real UI screens from the current build.
Quick Translate — fast snippet translation.
File Translation — upload & translate large documents.
Models — manage local AI models & hardware profile.
Glossaries — create groups for book-specific terminology.
Contact / Support
Send feedback, bug reports, or feature requests. This will create a ticket.
Destination: support@catscribetranslate.com
Newsletter
Subscribe to receive news and updates.
What happens next?
Your message is sent to our support inbox and stored as a ticket for tracking.
Tip: include your platform and app version. If you can, add steps to reproduce.
For privacy: do not paste passwords or sensitive info.
About
CatScribe Translate is an offline-first translation workspace built for large books and long documents. It focuses on consistency (names, places, terminology), robust chunked translation, and privacy.
What is a CAT tool?
CAT stands for Computer‑Assisted Translation. It helps translators work faster and more consistently using glossaries, translation memory concepts, segmentation/chunking, and QA workflows.
Offline-first
Your files stay on your machine. You can translate with local engines (LibreTranslate / MarianMT) and optionally use cloud providers only when you choose.
Built for books
Designed for fantasy series and long novels: glossary groups per book, consistent term usage, and stable long-running translation jobs.
Privacy & telemetry
Telemetry is anonymous and optional. It helps us understand stability and performance (crashes, speed, feature usage) without collecting the content of your documents.