Blog → podcast
Turn an article into a natural two-host conversation, generated in one pass with consistent voices.
Where it fits
Long-form, multi-speaker synthesis unlocks workflows that single-utterance TTS struggles with — and the 2026 recognition builds add live and on-device ones.
Educational dialogue
Teacher–student scripts, interviews and role-play voiced for courseware.
AI agent voice
Low-latency voice interface for agents (Realtime).
Corporate e-learning
Scale training narration without a studio.
Long-form narration
Narrate technical docs and long articles in a single coherent take.
Multi-speaker narration
Keep up to 4 voices consistent inside one model.
Live captions & meeting notes
Streaming ASR labels who is speaking as the meeting happens — no waiting for the recording to end.
On-device & offline
The 1.58 GB CPU build keeps sensitive audio on the machine.