LiveKit's video on Prism: Your voice agent just said the wrong thing. Here's how to stop it.
What you’ll need
- A Rime API key. If you don’t have one, sign up for Rime to create one. With LiveKit Inference, you can skip it.
- A LiveKit Cloud project, with its URL, an API key, and the key’s secret. Your own LiveKit server works too.
- uv 0.9.26 or later, which installs Python for the project if needed. The Rime SDK needs Python 3.11 or later.
- A browser with a microphone.
Build the agent
1
Create the project
Create a Python project:Install LiveKit Agents and the Rime plugin with its
realtime extra:Use LiveKit Agents 1.7.1
Use LiveKit Agents 1.7.1
To keep Agents 1.7.1, install the Rime plugin from this pinned commit in LiveKit’s official repository instead of the PyPI package.Add this setting to your project’s Install the plugin with Agents 1.7.1:
pyproject.toml so uv keeps the published Agents core:pyproject.toml
2
Set your credentials
Create
.env.local in your project folder with your own values. Your LiveKit Cloud project’s settings show its URL and let you create an API key and secret:.env.local
3
Save the agent
Save this script as Three parts control the conversation:
agent.py in your project folder. It registers an agent named prism-agent, which starts a Prism session in the aurora voice and greets the caller in each room it joins.agent.py
rime.realtime.RealtimeModel: connects the agent to Prism and selects its voice.@function_tool: lets Prism request the cafe’s opening hours from your code.generate_reply(): asks Prism to greet the caller first.
4
Start the agent
Run it in development mode from your project folder:The agent connects to your LiveKit project and logs
registered worker with "agent_name": "prism-agent". Leave it running. You can ignore the warning that dev is deprecated, or use lk agent dev from the LiveKit CLI instead.5
Talk to the agent
- In LiveKit Cloud, open your project’s Agent Console.
- Set the agent name to
prism-agent. - Select Start a session.
- When your browser asks, allow microphone access.
- The agent never joins: check that the console’s agent name is
prism-agentand the terminal showsregistered worker. - The agent joins but stays silent: look for a Rime plugin error in the terminal, and look up its code in Errors and limits.
How Prism changes a LiveKit agent
Prism handles the whole conversation. Keep these constraints in mind when adapting the example:
For settings, tool restrictions, retries, playback, and history, use the LiveKit plugin reference.
To choose your own transcription and language model, use the LiveKit TTS quickstart.
Use LiveKit Inference
You can also reach Prism through LiveKit Inference. It signs in with your LiveKit API key and secret instead of a Rime key, and LiveKit Cloud bills the usage. In the agent above, replace the line that creates the model:RIME_API_KEY out of .env.local. LiveKit’s Rime Prism page lists the Inference options.
Use your own LiveKit server
On your own server, which has no Agent Console, send the agent to a room with the LiveKit CLI:-
In
.env.local, setLIVEKIT_URL,LIVEKIT_API_KEY, andLIVEKIT_API_SECRETto your server’s values, and start the agent the same way. -
Send the agent to a room. Pass your server details with
--url,--api-key, and--api-secret, or set the same three variables in your shell. -
Join the
prism-testroom from any LiveKit client with a microphone, such as an app built with LiveKit’s frontend SDKs.
Next steps
Now that your agent answers, you can:- Take phone calls with LiveKit telephony.
- Write instructions that Prism follows, starting with rules for speech.
- Choose a voice that suits your callers.
- Configure the LiveKit plugin for settings, tools, retries, playback, and history.
- Move an OpenAI Realtime agent on LiveKit to Prism.

