Skip to content
Pi
0 bookmarks
Visit

Summary

Minimal, MIT-licensed terminal coding agent that adapts to your workflow via TypeScript extensions.

Screenshots

Description

Pi is a deliberately minimal terminal-based coding agent built around the idea that a harness should adapt to a developer's workflow instead of forcing one on them. Rather than shipping a large built-in feature set, Pi keeps only a handful of core tools and pushes everything else, sub-agent-style behaviors, plan modes, custom commands, into a TypeScript extension system that anyone can write, publish, and share as an npm or git package.

It talks to more than 15 model providers, including Anthropic, OpenAI, Google, Azure, Bedrock, Mistral, Groq, and Hugging Face, with mid-session switching between models and providers rather than a hard lock-in to one vendor. Sessions are tree-structured, supporting branching and compaction so a developer can explore an alternate approach without losing the original conversation thread, and Pi runs interactively, in print/JSON mode for scripting, over RPC for process integration, or embedded via SDK inside another application.

Created by Mario Zechner (creator of the libGDX game framework) and released under the MIT license with no SaaS backend, Pi is aimed at developers who want a fast, hackable, fully local coding agent they can extend with their own skills, prompt templates, and themes instead of working around a vendor's fixed feature set.

Reviews

Similar App Suggestions

Free, open-source local MCP gateway: register each server once, share it with every AI client, and cut tool-definition tokens by up to 91%.

Free

Agentic IDE that runs coding agents, self-healing QA tests and production monitoring in one workspace so fixes reach a PR automatically.

Freemium

An AI on-call agent that drops read-only probes into running production code, capturing live variable snapshots and stack context without a redeploy or a restart.

Freemium

A hosted LLM gateway from ngrok: one base URL and one key for public providers, your own keys and self-hosted models, with failover, scoped access and cost analytics.

Paid

Cloud-hosted long-horizon AI agents: pick a harness like Claude Code or Codex, pick a model, and reach the same agent from Slack, WhatsApp, iMessage, the web or a CLI.

Freemium