Skip to main content
A Pi pod runs the Pi coding agent 24/7 on its own machine. Pi is small and fast, and you extend it with skills and MCP tools. It is the third Agent Pod engine, next to ClawPod (OpenClaw) and HermPod (Hermes).

What you get

  • The Pi coding agent, always on, in your own pod.
  • Its own crypto wallet (Base and Solana), available to the agent as native tools, under the same spend controls as every pod. See Pod Wallet.
  • Preinstalled skills: x402-compute and x402-layer, plus Singularity marketplace tools, wired to your account with capped, revocable credentials.
  • Chat anywhere: from the dashboard, over the OpenAI-compatible API, or in your Telegram DMs.
  • Optional headless browser for reading and using web pages.
  • Sandbox access: run commands, edit files and open ports with the Sandbox API, the same as other engines.

Choosing an engine

ClawPod is the default when you do not name an engine.

Deploy a Pi pod

In the dashboard, open Agent Pods → Deploy a Pod and pick Pi pod. The rest of the flow is the same as any pod: see Deploy a Pod. With the Pods API, set engine to pipod:

Limits today

  • Singularity Layer models on a BYOK pod and BYOK fallback providers are ClawPod only for now. A BYOK Pi pod uses one provider: yours. A managed Pi pod runs Singularity Layer models as usual.
  • Pi chats in Telegram DMs. Running Telegram groups is a ClawPod feature.
The live list of engines, their status and features is in the pod catalog: GET https://compute.x402layer.cc/pods/catalog.