Chat will not answer: what to check

A model that never replies is usually Ollama not running, a model not pulled, a key that was not accepted, or an image sent to a model that cannot see images.

Last checked against the app on · View as Markdown

Steps

  1. Look for the engine banner at the top of the window. If it says the engine is not running or is restarting, deal with that first.
  2. If you are using Ollama, open Settings, then Intelligence, then API keys. The Ollama row shows the address the app expects and a live status line.
  3. If the status says Ollama is not running, install it from ollama.com and start it. If it says nothing answered at the address, start Ollama or choose another route. If it says no model is chosen, pick one in Settings. If it says the model is missing, pull that model in Ollama.
  4. If you are using a cloud provider, check the provider appears under "Your keys". A key that was saved with "Saved, but not checked" may still work, so send one message before replacing it.
  5. If you attached an image, check the model can see images. Otherwise remove the image or choose a model with vision.
  6. Choose a different model from the model button and send another message to tell a model problem from a setup problem.

How to tell it worked

A short message in Chat is answered.

If it does not work

The message says "This model cannot see images. Remove the image, or switch to a model with vision."
Do exactly that. The send button stays disabled while an image is attached to a model that cannot see it.
The message says the model has not published whether it can see images.
The image was not sent. Remove it, or switch to a model with known vision support.
The turn ends with an error you do not understand.
Report it from the bug icon or Ctrl+Shift+B and tick "The last error", which attaches the last error message the app recorded, message only, never a stack or a file path.