Skip to main content
PyAI ships a Model Context Protocol server, @pyai/mcp, that gives AI coding agents first-class PyAI tools: they can mint a free sandbox key and call Speak (TTS), Hear (STT), and the voice catalog directly instead of guessing endpoints.

​
Add to Cursor — one click

Add to Cursor Click the button — Cursor decodes the install config and prompts you to confirm. That’s the whole setup: no clone, no build, no key required to start (the agent can call create_sandbox_key for a free pyai_test_ key on demand). Prefer to do it by hand? Add this to .cursor/mcp.json (project) or ~/.cursor/mcp.json (global): 
{
  "mcpServers": {
    "pyai": {
      "command": "npx",
      "args": ["-y", "@pyai/mcp"]
    }
  }
}
To pin a key instead of minting a sandbox one, add "env": { "PYAI_API_KEY": "pyai_live_..." }.

​
Other agents

claude mcp add pyai -- npx -y @pyai/mcp
# or pin a key:
claude mcp add pyai --env PYAI_API_KEY=pyai_test_... -- npx -y @pyai/mcp

​
What the agent gets

ToolNeeds a key?What it does
create_sandbox_keyNoMint a free pyai_test_ key (no email/card), adopted for the session
get_startedNoCurated quickstart (auth, SDK install, TTS/STT/realtime snippets)
whoamiYesIntrospect the key — org, env, scopes, credit posture
list_models, list_voicesYesCatalogs
synthesize_speechYesText-to-speech to a file
create_transcription_job / get_transcription_jobYesAsync speech-to-text
The contract behind these tools is always https://api.pyai.com/openapi.json — the agent never invents endpoints.