# Pi 1.0 makes Codemode and MCP part of a stable minimal coding-agent harness Earendil has shipped Pi 1.0 as a stable, MIT-licensed coding-agent harness, adding Codemode with native MCP support, virtual models and deferred tool loading while keeping long-running agents in a separate experimental package. Pi’s first stable release is interesting less for another coding-agent version number than for what its deliberately minimal core now considers mature enough to include: MCP, code-driven tool orchestration and model routing. - Status: Active - Published: 2026-10-04T09:23:13+13:00 - Updated: 2026-10-04T09:23:13+13:00 - Categories: Artificial Intelligence, Web Development, AI Agents, AI Coding, Developer Tools - Tags: Coding agents, Earendil, Model Context Protocol, open source, Pi - Canonical HTML: https://beyondthe.news/dossiers/pi-1-0-codemode-mcp-stable-minimal-coding-agent ## What changed On October 1, 2026, Earendil released Pi 1.0, the first stable release of its MIT-licensed coding-agent harness. The core now includes Codemode, which gives Pi native MCP support and lets it work with non-LLM models such as decision and image models; extension support for virtual models; deferred tool loading; cache warming for Anthropic models; and transcript-aware mid-conversation system messages. Earendil says hundreds of thousands of people use Pi each week. It also released Pi Durable separately as an experimental package for longer-running agent applications rather than expanding the terminal agent itself into a larger orchestration system. ## Why it matters Pi is a useful counterpoint to coding agents that accumulate planning modes, subagents and built-in workflow layers. Its 1.0 boundary shows a smaller harness deciding that MCP and code-driven tool orchestration are now foundational enough to belong in the core while more opinionated long-running orchestration stays outside it. For builders, the practical appeal is an open, extensible substrate that can route across providers and models without tying the workflow to one vendor. The caveat is that Pi’s minimalism also pushes more responsibility onto extensions and the surrounding execution environment; stable 1.0 does not make arbitrary agent tools safe. ## Codemode moves tool orchestration into code Pi 1.0 includes Codemode with native MCP support and support for non-LLM models. Instead of exposing every integration as a large flat tool surface, the agent can write code that invokes tools and composes results. Earendil also added deferred tool loading to keep the active tool surface smaller. ## Virtual models make routing an extension concern Extensions can expose virtual models that combine underlying providers or decision models. Earendil demonstrates a router that plans with Claude Opus, uses Jev to detect the transition to implementation and then hands work to GPT-6 Luna. ## Long-running agents stay outside the core Earendil launched Pi Durable on the same day as an experimental package for longer-running agent applications. Keeping it separate preserves the terminal agent's smaller core while giving developers a path to build persistent or differently surfaced agent systems. ## Stable does not mean sandboxed Pi is designed as a malleable agent harness rather than a restrictive execution environment. Builders still need to reason about the permissions granted to tools, extensions, shells and MCP servers instead of treating the 1.0 label as a security boundary. ## Key details - Pi 1.0 shipped October 1, 2026. - The project is MIT licensed. - Earendil says hundreds of thousands of people use Pi each week. - Codemode adds native MCP and non-LLM model support. - 1.0 adds virtual models, deferred tool loading, Anthropic cache warming and mid-conversation system messages. - Pi Durable launched separately as an experimental long-running-agent substrate. ## Builder takeaways - Pi is worth evaluating when you want an open coding-agent harness without adopting a large built-in workflow system. - Use virtual models when routing or model composition belongs in your application rather than in a provider-specific product. - Treat MCP servers and extensions as executable authority and review their permissions accordingly. - Keep Pi Durable separate in your evaluation: it is experimental even though the terminal harness has reached 1.0. ## What to watch - Whether Pi 1.x keeps its minimal core as MCP and agent-extension ecosystems expand. - Adoption and stability of Pi Durable for persistent agent applications. - Independent measurements of Codemode token and latency savings across real tool workloads. - Security and permission tooling around extensions, MCP servers and code execution. ## Uncertainties - Earendil's weekly-user figure is first-party and has not been independently verified. - The release announcement does not provide independent benchmark data for Codemode efficiency or task quality. - Pi Durable is explicitly experimental and may change substantially. ## Sources - [Pi 1.0](https://earendil.com/posts/pi-1-0/) — Earendil · primary · 2026-10-01T00:00:00+13:00. Primary release announcement, feature list, licensing and first-party usage claim. - [The free and open-source AI coding tool 'Pi 1.0' has been officially released](https://gigazine.net/gsc_news/en/20261002-pi-1-0/) — GIGAZINE · secondary · 2026-10-02T00:00:00+13:00. Independent release coverage.