Test Voice AI in Your Browser

Browser testing uses your device's microphone (laptop, phone, tablet) to send audio directly to our voice AI pipeline. The AI processes it just like a phone call - using your configured STT, LLM, and TTS - and returns audio through your browser speakers. You see real-time transcripts and latency metrics.

Very close. The AI processing is identical - same STT, LLM, TTS, and prompts. The difference is audio quality: browser audio is typically better quality than phone audio (no telephony compression). So if something works in browser testing, it will work on phone. The reverse isn't always true - always do final testing with actual phone calls.

Any modern browser with microphone support: Chrome (recommended), Firefox, Safari, or Edge. Mobile browsers work too. You'll need to grant microphone permission when prompted.

Yes. Switch your agent's language settings and test any of the 24 supported languages directly in the browser. This is especially useful for testing Indian languages like Hindi, Tamil, or Telugu without needing local SIM cards.

The latency shown is end-to-end from when you finish speaking to when the AI starts responding. This includes STT, LLM, and TTS processing time. Note that browser audio has slightly lower latency than phone audio (no telephony network), so real phone latency will be ~100-200ms higher.

Yes. All agent capabilities work in browser testing, including function calling, CRM lookups, calendar bookings, and custom API integrations. You can verify that functions trigger correctly and see the full execution in debug mode.

Browser testing uses the same per-minute pricing as phone calls - you're billed for the STT, LLM, and TTS usage. There's no telephony cost since it's not going through phone networks. For most development testing, costs are minimal.

Yes. Each team member can run browser tests from their own device. Tests don't interfere with each other. You can share test recordings and transcripts for review.