Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.oneinbox.ai/llms.txt

Use this file to discover all available pages before exploring further.

The loop

On every call, audio cycles through three stages:
StageYou configure
STTProvider, model, language
LLMModel, system prompt, tools, knowledge bases
TTSProvider, voice, speed
OneInbox handles streaming, turn-taking, and call control.

What OneInbox runs for you

  • Real-time audio between STT, LLM, and TTS
  • End-of-speech detection
  • Silence timeout and end phrases
  • Transcripts and call metadata
  • Telephony routing for phone calls
  • Web call session tokens (server_url, participant_token)
You do not run WebRTC servers or SIP infrastructure yourself.

Call types

TypeWhat it isGuide
Web callAPI session via POST /v1/calls/webWeb calls
Web SDKBrowser client that joins a web callWeb SDK
Outbound phoneAgent dials a numberPhone calls
Inbound phoneCaller reaches your numberPhone calls

One agent, many calls

Create once. Pass per-call context with variables — no need to clone agents.

Next steps