HostDesk Documentation

Welcome to the HostDesk voice agent development platform. This documentation covers how to get a phone number for your agent and how to use the SDK to build and run voice AI in minutes.

Note on naming: the product is HostDesk. The Python package is published as echoai and the API key environment variable is ECHOAI_API_KEY — those are the original package identifiers and are unchanged, not a separate product.

Quick links

• Acquire a phone number — Rent and configure a number for your voice agent
• Use the SDK — Install the SDK, handle inbound/outbound calls, and run your agent
• Assistants API — Create and manage voice assistant configurations
• Function Calling — Webhook-based tool use during calls
• CLI Commands — Scaffold projects, listen for events, run workers
• Web/React SDK — Add voice AI to web apps with JavaScript/React
• Webhook Signatures — Verify webhook requests with HMAC-SHA256
• Call Routing — Transfer calls and hold/resume

What is HostDesk?

HostDesk gives you a low-latency voice pipeline (STT, LLM, TTS), voice on web and phone, and a dashboard to configure your agent. You can:

• Connect in the browser — Use the Playground to talk to your agent with live transcript and barge-in
• Receive or place phone calls — Attach a phone number to your agent; place outbound calls via the API or SDK
• Configure agent, knowledge, and tools — Use the Agent page to set your backend URL, knowledge base, and tools

Get started by acquiring a phone number and then integrating the SDK.