#Offline Translation

Offline translation is for workflows where translation should run locally and remain usable without relying on a live cloud service.

Offline-first does not mean every setup is identical. Your available engines, models, speed, and quality depend on what is installed and what your hardware can run comfortably.

#Benefits

• Work can continue without a stable internet connection.
• Local files stay in your controlled workspace.
• Performance is more predictable once your local setup is working.
• You can test and tune your workflow before translating a full project.

#Offline Workflow

1. Install or enable a local translation engine supported by your setup.
2. Confirm the engine works with a short test.
3. Import your source file.
4. Add glossary terms for important names and terminology.
5. Translate in batches.
6. Review in the CAT Editor.
7. Refine only the segments that need extra work.
8. Export and inspect the result.

#Performance Expectations

Offline translation uses local CPU, GPU, RAM, and storage. Larger files and larger models can be slower. If translation feels too slow, reduce batch size, use a lighter profile, or translate by chapter.

See Performance Guidelines for practical advice.