Pacerelle docs
Everything you need to create an account, connect a local agent, and build safely with Pacerelle.
Pacerelle lets you chat with local AI agents from the browser. These docs are written for users and agent builders who use the Pacerelle website, apps, and SDKs.
01Start from zeroCreate an account, create an agent, copy credentials, and send the first message.02Set up credentialsFind the agent id and token, store them safely, and know when to rotate them.03Build the runtimeUse JavaScript, Python, or MCP to receive messages and send replies.
Recommended reading order
QuickstartFollow the shortest complete path from signup to a working echo agent.
Account and device setupLearn what happens in the UI when you sign up, sign in, or use another browser.
Agents and credentialsLearn where to get
PACERELLE_AGENT_ID, how the one-time token works, and how to keep it safe.SDK integrationConnect a real local process with Node.js, Python, or your preferred workflow.
Screens you will use
EmailUsed to sign in
Display nameShown in the app
Password12 characters minimum
NameDev assistant
Agent idCopied to
.envTokenShown once
What this documentation covers
- User setup: signup, login, trusted browser behavior, and account recovery expectations.
- Agent setup: creating agents, copying credentials, checking online status, and rotating tokens.
- Runtime setup: JavaScript SDK, Python SDK, MCP server, local environment files, and process startup.
- Product behavior: conversations, attachments, widgets, safe approvals, and troubleshooting.
- Production readiness: secret handling, process supervision, token rotation, and safe agent behavior.
Start with Quickstart if you want the fastest working path.