Run the AI chat on your own computer with Ollama or LM Studio. Nothing goes to a cloud unless you choose a cloud model.
Free during launch. No account, no watermark.
The AI chat runs on your PC through Ollama or LM Studio. No key, no upload.
Run Ollama or LM Studio
Load a model that can use tools, then pick it in the AI chat's settings.
Ask in plain words
The model gets a live summary of your project with its real clip IDs, and only the tools your request needs, so the prompt stays small.
Check what changed
A checklist shows each edit, written by the app from the project, not by the model. One undo reverts the turn.
Free the graphics card
Before an export or captions, Evaloon unloads the model so the work has the memory it needs.
Does it work offline?
Yes. With a local model the chat runs on your PC and needs no internet connection.
Can a local model do everything?
Not quite. It does captions, silences and caption templates, and says what it skipped. Smart cut, picking graphic moments by meaning and skill scripts need a cloud model.
How do I know which model is answering?
The chat header always shows Local or Cloud and the model name.
Which models can I use?
Any model in Ollama or LM Studio that supports tool calls. A model without tools can chat, but can't edit.