Use Ollama (local AI)

Ollama is a free app that runs AI models on your own computer. With it, the AI Editor works with no key, no account and no per-use cost, and your project stays on your PC.

Set it up

1. Install Ollama from ollama.com and start it. It runs in the background.

2. Download a model that can use tools. Open a terminal (Windows Terminal or PowerShell) and run:

   ollama pull qwen3:8b

   qwen3:8b is the model Evaloon Editor suggests first. Other models that support tools also work.

3. In Evaloon Editor, open Settings, choose the AI Editor tab, and set Provider to Ollama (local).

4. Leave Base URL as http://localhost:11434 (Ollama’s usual address), then choose your model from the Model list. The list shows the models installed in Ollama.

5. Wait for the status line under the model. It should say Ollama is ready and name the model. Now open the AI Editor panel and try a request.

The URL and model are remembered in Settings.

What the status line can tell you

Message	What to do
Ollama isn’t reachable at …	Start Ollama, then select Check again.
… isn’t installed	Run ollama pull with the model name shown, or pick a model that is installed.
… doesn’t support tools, so it can’t edit your timeline	Choose a model that supports tools.
Choose an installed model	Pick one from the Model list.

Tips

• Bigger models are smarter but need more video memory. If edits are slow or fail, try a smaller model. Exact memory guidance is to be confirmed.
• Local models are good at templates, captions and removing silences. For cutting by what is said, or writing custom animation, a cloud model such as Claude does better: see Cloud keys.
• If you also use LM Studio, only one local model can sit in video memory at a time on a small PC. Unload one before using the other.

Note

The Base URL does not have to be your own PC. If you point it somewhere else, the app warns you that your project details and requests go to that server.