Initializing...
Line is cold. Start a call when you want to talk through the operation.
Initializing...
Live microphone transcription runs through the configured Mistral realtime model.
Loading available LLM models...
Browser built-in uses the device speech engine. Google Cloud uses the configured API voice.
Continuity memory plus the live decision context that the LLM sees before choosing tools. Tool results still appear separately in the tool-data popup.
Waiting for memory...
Exact tool payload captured from the live tool call.
Waiting for tool data...
Quit this run and clear the saved session from this device?