Aisha Lab
Where new voice AI gets tested.
Capabilities we are building and refining before they reach production. Everything here runs on the real Aisha API - try it yourself in Space.
Live
Emotional speech
Uzbek TTS with moods - Neutral, Cheerful, Happy and Sad - plus speed control from 0.5× to 2×.
Live
Speaker diarization
Speech-to-text that separates who said what, built for call recordings and meetings.
Live
Real-time voice agent
An agent that answers live phone calls, books appointments and hands off to humans when needed.
Beta
Meeting intelligence
Automatic transcripts and structured summaries for meetings in Uzbek, Russian and English.
Aisha Lab
The voice orb
The same real-time visualizer that powers our agent interfaces. It reacts to speech, listening and silence.
Try every experiment in Space
One account for the currently available models and agents.