Quick Start

Know if your voice agent is actually working. Parlot’s MIT parlotize() sidecar instruments LiveKit Agents — Cloud, self-hosted, or LiveKit on Telnyx — and models each call as turns: timeline, multi-agent graph, and goal completion with an evidence trail to the proving turns and audio.

Install

# Using pip
pip install parlot-instrumentation-livekit

# Or using uv
uv add parlot-instrumentation-livekit

# Or via the parlot meta-package
pip install "parlot[livekit]"

Instrument

Call parlotize("…") before constructing AgentSession, then await ctx.connect() before session.start():

from parlot.instrumentation.livekit import parlotize

parlotize("my-agent")

from livekit.agents import AgentSession, JobContext, WorkerOptions, cli


async def entrypoint(ctx: JobContext):
    await ctx.connect()

    session = AgentSession(...)
    await session.start(agent=..., room=ctx.room)

Environment

export PARLOT_ENDPOINT=https://ingest.parlot.ai
export PARLOT_API_KEY=<org-scoped-key>

Mint the API key in Parlot Settings → API Keys. New tenants record session audio by default (Settings → Recording allowlist *); clear the allowlist or use parlotize(record=False) to disable. Audio uploads to R2 and confirms when you open the session (lazy R2 HEAD reconcile).

Next

• Core concepts
• LiveKit guide
• Python API Reference
• TypeScript Reference
• Already running LangGraph / LangChain alongside voice? See the LangGraph guide.

License

Instrumentation is MIT. The companion analysis platform is BSL 1.1: free production use up to 4,000 turns/month for your own agents and agents you operate for clients; above that, or to offer Parlot as a hosted service, requires a commercial license. Each version converts to Apache 2.0 four years after its first public release. Enterprise governance features require an EE license key.