

Pi Agent is a minimal, terminal-native coding agent from Earendil Inc., hosted at https://pi.dev/, that treats the harness as something to be shaped by the developer rather than a fixed product. Pi Agent runs in the terminal as a TUI, supports a tree-structured session history, ships skills and AGENTS.md/SYSTEM.md context files, and is unusually token-efficient because of its minimal default system prompt. The tagline is "There are many agent harnesses, but this one is yours."
Four runtime modes are bundled. Interactive mode is the full TUI with rich keyboard shortcuts and a themable interface. Print mode runs Pi Agent headlessly with a single query for shell scripts. JSON mode streams structured events from a run for piping into other tools. RPC mode exposes Pi Agent over JSON on stdin and stdout, letting non-Node processes drive it. An SDK lets teams embed Pi Agent inside their own applications; OpenClaw is cited as a real-world example of the SDK in production.
Pi Agent is model-agnostic. It works with Anthropic, OpenAI, Google, Azure, Bedrock, Mistral, Groq, Cerebras, xAI, Hugging Face, Kimi For Coding, MiniMax, NVIDIA, OpenRouter, Ollama and more. Authentication can use API keys or OAuth. The /model command and Ctrl+L shortcut switch models mid-session, Ctrl+P cycles through favourites, and new providers can be added through models.json or extensions.
Context engineering is central. Pi Agent loads AGENTS.md instructions from ~/.pi/agent, parent directories and the working directory, supports per-project SYSTEM.md overrides, auto-summarises older messages via compaction, and ships a skills system for capability packages loaded on demand. Prompt templates are reusable Markdown prompts invoked by typing /name. Extensions, written in TypeScript, can inject messages before each turn, filter history, implement RAG or build long-term memory. Sessions are trees; /tree navigates, /export writes HTML, /share pushes to a GitHub gist for a shareable URL.
Pi Agent is for developers who want maximum control over their coding workflow.
ADHD Test - an educational self-screen built on the WHO/Harvard ASRS v1.1 screener.
Free PSL Rating delivers a private, in-browser PSL attractiveness score with a photo confidence rating.
AI-Powered Facial Attractiveness Evaluation
AI-powered impromptu speaking practice — from random topic to real feedback.
Accurate PDF to Markdown Conversion