Ollama’s new “launch pi” command lets you spin up a coding agent in seconds
Ollama added a one-command way to launch a minimal coding agent. It is called Pi, and it powers OpenClaw under the hood.
Ollama added built-in support for Pi in version 0.17, letting you launch a terminal-based coding agent with a single command.
ollama launch pi --model kimi-k2.5:cloudPi is the minimal coding harness that powers OpenClaw under the hood. It ships with four basic tools: read, write, edit, and bash. That is it. Everything else you add yourself.
- Extensions: Add capabilities through TypeScript plugins.
- Skills: Drop in specialized behaviors for specific workflows.
- Prompt templates: Customize how the agent communicates.
- Model support: 15 plus providers, local via Ollama or cloud models like Kimi K2.5.
The design philosophy is deliberately lean. Pi gives you the plumbing and gets out of the way. Users have been comparing it to building with Lego — you assemble what you need rather than fighting a framework that assumes it knows your workflow better than you do.
If you want the full OpenClaw experience with persistent memory, web search, and a dashboard, that is still available via ollama launch openclaw. Pi is for developers who want the coding core without the rest of the stack.
The community reaction has been positive, with a few honest caveats. It works well for personal and custom workflows. For production use, code review before any merges is still on you.
If you need on-demand GPUs for training, fine-tuning, inference, or running open-source models, give RunPod a try.
- Available hardware: H100, H200, A100, L40S, RTX 4090, RTX 5090, and 30+ more
- Cost: significantly cheaper than AWS or GCP, billed per second, no contracts
- Setup: spins up in under a minute, 30+ regions worldwide
Get the core business tech news delivered straight to your inbox. We track AI, automation, SaaS, and cybersecurity so you don't have to.
Just read what you want, and be done with it.
