Docs

Appearance

Voice Agent API

Voice agents that answer real phone calls. Phone or browser, from one API or any MCP client.

Get API key →
Quickstart
Open Studio →

Place a browser call ​

This request needs an API key and an agent id. It does not need a phone number. Connecting a carrier for a real phone call is the quickstart.

bash
curl -X POST https://api.rymi.live/v1/calls \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "agent_id": "550e8400-e29b-41d4-a716-446655440000",
    "participants": [
      { "transport": "webrtc", "identity": "browser-tab-1" }
    ]
  }'
typescript
import { Rymi } from '@rymi/node';

const rymi = new Rymi();
const call = await rymi.calls.create({
  agent_id: '550e8400-e29b-41d4-a716-446655440000',
  participants: [{ transport: 'webrtc', identity: 'browser-tab-1' }],
});
python
from rymi import Rymi

rymi = Rymi()
call = rymi.calls.create(
    agent_id="550e8400-e29b-41d4-a716-446655440000",
    participants=[{"transport": "webrtc", "identity": "browser-tab-1"}],
)
json
{
  "mcpServers": {
    "rymi": {
      "type": "http",
      "url": "https://mcp.rymi.live/",
      "headers": {
        "Authorization": "Bearer rymi_your_secret_key"
      }
    }
  }
}

What do you want to build? ​

Outbound calls
Dial up to 500 PSTN participants into one call in a single request. Built-in retry, health monitoring, variables.
Quickstart →
Inbound agent
BYOC a Plivo/Twilio/Vonage number and route incoming calls to a voice agent.
Integrate telephony →
Browser widget
Put a WebRTC voice agent on any site through the REST API, with publishable keys scoped to an agent.
Embed in browser →
AI agents via MCP
Let Claude, Cursor, or any MCP client create and manage Rymi voice agents through tool calls.
Set up MCP →
Post-call analysis
Signed webhooks with summaries, sentiment, and structured JSON extraction after calls.
Post-call intelligence →
Webhook events
HMAC-signed events for call lifecycle, intelligence-ready, and DNC flags.
Webhook reference →

Client libraries ​

Node.jsnpm i @rymi/nodePythonpip install rymiBrowsernpm i @rymi/webReactnpm i @rymi/reactMCP newmcp.rymi.live

Recent updates ​

May 2026
Multilingual agents. Select the languages your callers speak and Rymi resolves a full speech-to-text, reasoning, and voice stack for each. Languages →
May 2026
Knowledge sources. Upload your docs and agents answer from your own content with full-text or embedding retrieval. Knowledge sources →
Apr 2026
MCP server. Any MCP-capable AI client (Claude Desktop, Claude Code, Cursor, or a Rymi voice agent itself) can now create and manage voice agents. Read the docs →
Apr 2026
Flat-diff change-set API. Validate and apply typed config updates via POST /v1/agents/apply-changes. See reference →
Apr 2026
AI-drafted agents. Generate a full voice agent config from a one-line description with POST /v1/agents/generate.