Talk. Droski thinks. Droski speaks.
Voice capture, governed reasoning, and audible output are separate on purpose. Your microphone is off until you choose to use it.
Ready
Current mode
General Droski
No private matter is loaded by default.
Ready. Tap Start talking or type below.
Voice input does not authorize a filing, message, payment, external action, verified fact, or durable-memory decision.
Conversation
One governed Droski thread.
Your transcript goes to the same Droski reasoning route as typed input. ElevenLabs is the audible output identity, not a second reasoning brain.