pi

This article explains how compaction works in the Pi coding agent, summarizing old conversation history when the context window nears its limit, and details Pi's specific implementation and triggers.

The author criticizes Claude Code's increasingly large system prompt (32k tokens) for degrading cost, latency, and performance, and praises Pi's minimalist 1k-token approach with plugins as a better philosophy for coding agents.

A comparison article pitting AI coding tools Claude Code, OpenCode, and Pi against DeepSeek V4 Flash in a harness showdown.

Introduces several noteworthy extensions in the Pi ecosystem, including a safe discussion mode, web search, multi-agent workflow orchestration, and a model usage visualization tool.

LazyPi offers a one-command setup for Pi coding agent that installs 60+ community skills, 67 themes, MCP support, sub-agent support, and more, providing a curated starting point to avoid research and configuration overhead.

An exploration of why two statistically identical sequences—pure noise and the digits of π—differ in compressibility, distinguishing between statistical redundancy and process-based compression.

Newer Anthropic models like Opus 4.8 and Sonnet 5 are worse at using third-party editing tools (e.g., Pi's) compared to older models, likely because they were trained to use Claude Code's built-in edit tool via RL, causing them to invent extra fields in tool calls.

pi-workflow is a workflow orchestration tool designed for the Pi AI assistant, enabling it to execute multi-step, repeatable workflows, including deep research, code review, specification check, and impact assessment. It defines workflows as JSON stage graphs and supports multiple execution modes.

The author shares their experience using various AI coding assistants (Claude Code, Codex, Pi) for code review and refactoring, finding frontier models surprisingly effective at catching subtle bugs but noting the lower quality and erratic behavior of some tools.

A comparison of OpenAI's Agent SDK with other AI agent platforms including Hermes, Pi, and OpenClaw, analyzing their features and capabilities.

A guide on building a personal coding harness using Pi, Codex, and OpenRouter to achieve top frontier performance without heavy frameworks.

A guide for running Pi AI agent securely inside a Docker Sandbox while running llama-server on the host machine for local GPU inference.

BohuTANG found through quantitative analysis that adding dedicated grep and glob functions to the pi tool reduces the detours of LLM-generated bash and improves efficiency.

A deep dive into the architecture of Pi, a minimalist AI coding agent, explaining how it works.

In version 0.0.14, Pi now supports integrations with Codex and Claude Code.

A user describes how Qwen3.6 35B, combined with the 'pi' tool, has transformed their computer workflows, allowing natural language control of the OS and automated task execution. They successfully built a landing page from voice messages entirely locally, demonstrating the model's practical utility.

A pull request for llama.cpp fixes the constant prompt processing issue that occurs when using OpenCode or Pi with the library.

pi-subagents is a tool that allows Pi to delegate tasks to focused child agents for code review, implementation, and other workflows, enabling parallel and background jobs.