pi

pi is the terminal agent for the person who has looked inside the others and wanted less. It is one TypeScript monorepo, an LLM API, a loop, a TUI and the agent, and it deliberately leaves out MCP, so if your day depends on MCP servers this is the wrong tool on purpose. The trait that decides it is legibility: you can read the whole harness and change what you dislike.

Pick it if you want a harness you own and are comfortable writing a small extension. Pick OpenCode if you need MCP today, and Claude Code if you want a vendor's defaults and a single bill.

The risk is the extension model. Extensions are TypeScript modules that add tools, commands, events and UI to the running agent, and packages bundle them for distribution, so installing one runs somebody's code inside the process that holds your API keys and your shell. There is no sandbox around an extension, only trust.

Read the package before you add it, and keep the list short. What it does right: the official install is npm install -g --ignore-scripts, which shows the author applying the same suspicion to npm that you should apply to extensions.

The structure is layered and documented. 1. A unified LLM API abstracts the providers. 2. An agent loop sits on it. 3. A TUI and the coding-agent CLI sit on the loop, so each layer is usable without the one above. 4. Sessions are trees: a conversation can branch, be navigated, and be compacted, with abandoned branches summarised rather than dropped, which is a principled answer to context growth.

No benchmark is published and no capability claim is made that would need one. The observation: a session tree is version control for a conversation, and the rest of the board still uses a scrollback.

A corporation exists, Earendil Inc., and it owns the most-starred terminal agent on this board at 101,431, released in August 2025, with nothing for sale: bring-your-own-key, MIT, no hosted plan. That is either a very long game or a company whose business is elsewhere and treats the agent as reputation.

Pricing power: none while the product is free and the models are someone else's. Moat: the audience and the extension ecosystem, if it forms. Likely acquirer: a lab that wants a clean, small loop and a hundred thousand watchers. Position: watch the company, not the repo.

The demo is a small agent in a nice TUI. Procurement: /login signs the engineer into a subscription provider from the terminal, and sixty engineers doing that on personal plans is sixty consumer agreements applied to company code, which legal will not accept; API keys are the alternative and mean sixty keys. There is no SSO, no audit log, no admin and no vendor support line.

There is a headless mode, pi --mode json, which emits an event stream a pipeline can read; that is the one box it ticks. Onboarding is an npm install. Not yet; company-issued keys and a provider contract first.

MIT and TypeScript, and the extension API is the product: a module can add a tool, a slash command, an event handler or a UI component, and a package bundles extensions, skills and prompts so I can hand my setup to someone else. Local models are first-class through llama.cpp with /llama commands to manage them, and hosted keys go in environment variables like ANTHROPIC_API_KEY.

The absence I notice is a plugin registry with signatures, which the package system will need. A fork is a monorepo build. This is the harness I would write, already written.