MAI Realtime 1

Speech-to-speech model with natural conversational flow and precise steerability

Start a live voice conversation using your microphone. Sign-in is required.

Get consent for other voices
Don't share anyone else's voice without their permission.
Leave out anything sensitive
Your audio is processed only for this live conversation.
Your voice isn't used to identify you
Or anyone else who isn't a public figure.
For personal use only
Not for commercial, school or work use.

Welcome to MAI Labs!

MAI Labs is intended for personal, non-commercial use only. Do not use MAI Labs or generated content for any school or work-related activities.

By continuing, you agree to our Terms of Use (opens in a new tab) and Privacy Statement (opens in a new tab).

Listens, understands context and responds with human-like fluency.

A speech-to-speech model that pairs MAI’s audio understanding and natural voice with real-time reasoning, so any team can build steerable, responsive voice agents through one interface.

Big Bench Audio
0 %

Accuracy on spoken reasoning questions from the Big Bench Audio benchmark.

Less cost
$0.01

Per minute, with reasoning costs charged separately. At Input it's $0.01, and at output it's $5 per 1M tokens.

Time to first audio
0 ms

P50 Latency on BigBench Audio

Performance

Dig deeper

Register your interest

Our speech-to-speech model is in preview. Tell us what you’re building, and we’ll find the right next step for your team.

Register interest

Data & privacy

Manage your Microsoft AI model-service data.

English (United States)
Your Privacy Choices Opt-Out Icon Your Privacy Choices
Consumer Health Privacy Sitemap Contact Microsoft Privacy Manage cookies Terms of use Trademarks Safety & eco Recycling About our ads