Skip to content

Set the local Ollama model

Item Current behavior
Default llama3.1:8b
Valid values An installed model recognized by the bundled local Ollama runtime.
Existing sessions Existing transcripts and summaries are unchanged.
New actions Future Local Mode summaries, session chat, and global-chat answers use the selected model.
Storage synthezia-settings in webview local storage.
Privacy Requests go to the local loopback Ollama runtime. Pulling a model can contact its configured registry.
Restart None after a successful pull and inventory refresh.
  1. Select Local (Ollama) under AI Provider.
  2. Open Models and choose an installed model under Summarization & Chat.
  3. Pull another catalog model when necessary, then refresh the inventory.

An unavailable runtime or missing selected model stops a Local Mode action at a model requirement rather than silently using another provider.