Models are downloaded once and run locally. In Evano Studio you can install recommended models from the Models page — no terminal needed.

From the app

  1. Open the Models page.
  2. Pick a recommended model (e.g. Gemma, Qwen for coding, Llama for chat, Mistral, or a DeepSeek distill).
  3. Click Install and watch the progress. Models are large (often 2–5 GB) — make sure you have disk space and RAM.

Advanced (optional)

If you prefer the terminal, you can also pull a model directly with Ollama:

ollama pull gemma3:4b

Use the exact model tag (e.g. gemma3:4b) when configuring an agent — a family name alone may not match an installed model.

← Back to all docs