pi-mono: Create Your Own AI Coding Agent in Your Terminal
- GitHub Stars: 5.9k
- Language: TypeScript 96.5%
- License: MIT
Why This Project Is Popping Up
A developer felt Claude Code had become too complex. Mario Zechner experimented with LLM coding tools for 3 years and eventually decided to build his own tool.[Mario Zechner]
pi-mono is an AI agent toolkit built with the philosophy “don’t build it if you don’t need it.” It starts with a 1000-token system prompt and 4 core tools (read, write, edit, bash). Very lightweight compared to Claude Code’s thousands of token prompts. What’s in it?
- Integrated LLM API: Use 15+ providers including OpenAI, Anthropic, Google, Azure, Mistral, Groq in one interface
- Coding Agent CLI: Write, test, and debug code interactively in the terminal
- Session Management: Pause and resume work, branch like git
- Slack bot: Delegate Slack messages to the coding agent
- vLLM pod management: Deploy and manage your own models on GPU pods
- TUI/Web UI library: Build your own AI chat interface
Quick Start
# Install
npm install @mariozechner/pi-coding-agent
# run
npx pi
# or build from source
git clone https://github.com/badlogic/pi-mono
cd pi-mono
npm install && npm run build
./pi-test.sh
Where Can I Use It?
If Claude Code’s $200/month is too expensive and you prefer working in the terminal, pi could be an alternative. You only pay for API costs.
If you want to use self-hosted LLMs but existing tools don’t support them well, pi is the answer. It even has built-in vLLM pod management.
Personally, I think “transparency” is the biggest advantage. Claude Code runs invisible sub-agents internally to perform tasks. pi lets you directly see all model interactions.
Things to Note
- Minimalism is the philosophy. MCP (Model Context Protocol) support is intentionally omitted
- Full access called “YOLO mode” is the default. Be careful as permission checks are looser than Claude Code
- Documentation is still lacking. Read the AGENTS.md file carefully
Similar Projects
Aider: Also an open source terminal coding tool. Similar in being model-agnostic, but pi covers a broader scope (UI library, pod management, etc.). [AIMultiple]
Claude Code: Has more features but requires a monthly subscription and has limitations on customization. pi allows freely adding features through TypeScript extensions.[Northflank]
Cursor: AI integrated into an IDE. If you prefer GUI over terminal, Cursor is better.
Frequently Asked Questions (FAQ)
Q: Can I use it for free?
A: pi is completely free under the MIT license. However, if you use external LLM APIs like OpenAI or Anthropic, those costs apply. You can use it without API costs by running Ollama or self-hosted vLLM locally.
Q: Is the performance good enough to replace Claude Code?
A: In Terminal-Bench 2.0 benchmarks, pi with Claude Opus 4.5 showed competitive results with Codex, Cursor, and Windsurf. This proves the minimalist approach doesn’t compromise performance.
Q: Does it support languages other than English?
A: The UI is in English, but if the connected LLM supports other languages, you can communicate and code in that language. You can write code with prompts in any language by connecting Claude or GPT-4.
If you found this article useful, please subscribe to AI Digester.
References
- GitHub repository
- What I learned building an opinionated and minimal coding agent – Mario Zechner (2025-11-30)
- Claude Code vs Cursor Comparison – Northflank (2026-01-15)
- Agentic CLI Tools Compared – AIMultiple (2026-01-20)