Positional mix and orbiting TTS

September 2, 2026 · release

@voicethere/agent

Voice+Data sessions can route each client through a shared stereo mix: mix groups decide who hears whom, setClientPose moves listeners in space, and setTtsPose places each client’s TTS speaker relative to their ears.

What shipped

• Mix group APIs — createMixGroup, addClientToMix, removeClientFromMix
• Live poses — setPositionalMixing, setClientPose, named setDefaultMixPlacement
• Per-client TTS — setTtsPose(sessionId, pose) pairs with speak(sessionId, text); clearTtsPose and setTtsMixPlacement for fallbacks
• Dashboard starter Orbiting TTS (positional-tts) — TTS circles each listener on a 50 ms timer

Equal-power stereo panning with inverse-square distance (Y-up, look −Z). No HRTF.

Try it

• Docs: Positional mix & orbiting TTS
• Template catalog: Agent templates
• Dashboard: create a project → Template → Orbiting TTS, deploy with Voice+Data runner mode

npx @voicethere/agent build --entry templates/positional-tts/agent.ts

Similar posts

• Mute, mix status, and STT in positional mix
• Clip playback for voice agents
• Recording-consent agent template