Choose a voice integration

Choose who owns the conversation before choosing an audio transport.

Your situation	Integration	What you build
You already have a signed-in browser app, chat UI, and reply engine	Existing Concierge Voice	A token endpoint and signed HTTPS reply hook
You want ConverSimple to answer visitors on a website	Script or iframe widget	An agent, deployment, and embed snippet
You want a phone line answered by an agent	Twilio phone deployment	A connected Twilio number and agent binding
You want custom Python tool execution	Python SDK dispatcher	A long-running dispatcher and tool handlers

The ordinary website widget runs a ConverSimple agent. Existing Concierge Voice is different: it captures and transcribes speech, sends final utterances to your backend, and speaks the text your backend returns. It does not run a ConverSimple LLM or choose tools. The Python SDK is different again: it hosts customer-defined tools for a ConverSimple agent; it is not the browser voice SDK. If your workflow requires a live person to take over a phone call, review current limits before designing the experience. A flow handoff message is not a completed phone transfer.

Overview

First conversation

Choose a voice integration