Choose an exercise
Local AI examinerOff. Standard practice available.
Optional Qwen model runs on your device. First use downloads model files. Requires WebGPU and sufficient memory. Microphone remains required.
RoleplayStep 1
STEP 1 · LISTEN
Speak into your microphone to submit. The text is a read-only transcript.
Speaking coach
Only continue after checking your response for the best adaptive follow-up. Every submitted answer requires microphone recognition.