Voice Chat with AI
Commercial use OK400+ models54 voicesNo sign-up needed
Model:+ GPT-5, Claude, Gemini
Voice Conversation with AI
Tap the microphone to start. Speak naturally - AI will respond when you pause.
Tap to start conversation
Voice Settings
Advanced options
~200 tokens/exchange (STT + Chat + TTS)
Have a real-time voice conversation with AI for free. Speak, get transcribed, and hear AI responses aloud. No sign up required.
How to Use AI Voice Chat
1
Tap the microphone
Allow microphone access once and start talking. No account needed.
2
Speak, then pause
Your turn sends itself after about 1.5 seconds of silence. Prefer typing? Use the field under the mic.
3
Hear the answer
The reply comes back as text and is read aloud in the voice you picked. Free for personal and commercial use.
Use this tool via API
Automate this tool from your own code. OpenAI-compatible REST endpoint, Bearer-token auth, no extra SDK required. Token costs match the web interface.
AI Voice Chat FAQ
A free voice conversation with AI on Rewind.ai. You speak, your words are transcribed, the model answers, and the answer is read back to you out loud. No sign up needed to start.
Yes. You get 2,500 free tokens per day to use AI Voice Chat and every other tool on Rewind.ai. A free account raises that to 5,000 tokens/day. One spoken exchange costs about 200 tokens in total — transcription, the reply and the read-back together.
Three steps run per turn: your recording goes to speech-to-text, the transcript goes to the chat model, and the reply goes to text-to-speech. Your turn ends on its own after about 1.5 seconds of silence, so you do not have to press anything to send.
For speaking, yes — the page asks your browser for microphone access the first time you tap the mic. Without one you can still use the tool: type your message in the field under the mic and the answer is still read aloud.
54 voices across nine languages, from the open-source Kokoro model. Pick one under Voice Settings; the default is Heart, an American English voice.
Your data is processed on our servers and isn't stored permanently unless you choose to save it. We don't sell or share it. The Clear button drops the conversation from the page immediately.
Yes. The field under the mic sends a typed message through the same pipeline, and the reply comes back both as text in the transcript and as audio.
The voices cover American and British English, French, Hindi, Italian, Japanese, Mandarin Chinese, Portuguese and Spanish. The chat model itself understands far more languages than there are voices to read them back in.
Yes. Voice Settings has a Speed control with Slow, Normal and Fast. It changes how quickly the answer is read back, not how the model thinks.
Qwen 2.5 7B Instruct by default. The Model picker above the card switches to any of 400+ models, including GPT-5, Claude and Gemini — premium models cost more tokens per exchange. The read-back always uses Kokoro.
Within the conversation, yes: the last 20 turns travel with each new question, so you can refer back to what was said. Clear resets it, and nothing carries over to a new visit.
Any current desktop or mobile browser that can record audio and play it back — Chrome, Edge, Safari and Firefox all work. Microphone access has to be allowed; if you decline it, typing still works.