{"slug": "pi", "name": "Pi", "category": "ai-coding", "category_name": "AI coding tools", "description": "Open-source, minimal terminal coding agent harness that you extend with plugins and skills.", "recommenders": 223, "commenters": 358, "warners": 9, "score": 224.368, "rising": 91.65, "first": "2026-01-30", "latest": "2026-10-31", "why": "Loved for its minimal, lightweight design and easy extensibility, plus token-efficient context handling that works with both local models and Claude or GPT subscriptions.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 245, "neutral": 136, "compare": 26, "warn": 9}, "spark": [0, 0, 1, 11, 1, 12, 5, 42, 42, 48, 30, 53], "listed": true, "loved": ["Extensibility", "Lightweight minimalism"], "caveat": "Plugin quality", "short_quote": "I really like Pi and have recommended it to many people. It has been my daily driver for months", "icon": false, "domain": "pi.dev", "github_repo": null, "official_url": "https://pi.dev", "pricing": "open source", "aliases": ["Pi", "pi", "pi.dev", "pi coding agent", "pi agent", "Pi Coding Agent", "pi-mono", "Pi.dev", "Pi coding agent", "pi-agent", "Pi agent", "pi-dev", "Pi Mono", "pi-coding-agent"], "verdict": [{"text": "Loved for its minimal, lightweight design and easy extensibility", "cites": [{"id": 801584176, "item_id": 50099011, "url": "https://news.ycombinator.com/item?id=50099011", "by": "usernametaken29", "date": "2026-10-15", "stance": "recommend", "quote": "I recently switched to pi which uses an order of magnitude less RAM (hundreds vs dozens)", "excerpt": "I wish they\u2019d invest more time in performance improvements over code mods. I recently switched to pi which uses an order of magnitude less RAM (hundreds vs dozens) and noticeably less CPU on my machine\u2026 OpenCode is really nice but it\u2019s about as heavy as antigravity and is substandard in features\u2026", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 798849136, "item_id": 49928071, "url": "https://news.ycombinator.com/item?id=49928071", "by": "lo5", "date": "2026-10-01", "stance": "recommend", "quote": "It feels light and nimble.", "excerpt": "I use Pi with only a couple of extensions and using less and less of CC and Codex everyday. It feels light and nimble. Codemode from yesterday's release made it even better. I wish I could use my Claude subscription via Pi.", "context": "", "use_case": "AI coding agent", "switched_from": "Claude Code", "story": "Pi 1.0"}, {"id": 798881744, "item_id": 49930109, "url": "https://news.ycombinator.com/item?id=49930109", "by": "by2kaapi", "date": "2026-10-02", "stance": "recommend", "quote": "But loved the minimalism, extensibility! It's a very capable harness.", "excerpt": "Switched to pi to avoid having to keep skills/MCP configs in sync across agents like claude and codex to have consistent experience. But loved the minimalism, extensibility! It's a very capable harness. Have my own extensions for MCP, subagents, ui, history management etc. overall i feel my workflow has got a lot better. Love pi!", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798898528, "item_id": 49931158, "url": "https://news.ycombinator.com/item?id=49931158", "by": "eloisant", "date": "2026-10-02", "stance": "recommend", "quote": "I switched from OpenCode to Pi and never looked back.", "excerpt": "Yes, I switched from OpenCode to Pi and never looked back. What's great with Pi is that it's very easy to extend, because it knows its own doc. So you tell it \"implement a plugin that does this\" and it does it directly in its own folder. That means I have a plan mode that works exactly the way I want, I have a hook that cleans up added comments after each change, I can ask \"pull this github PR and assess the comments", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "Pi 1.0"}, {"id": 804599345, "item_id": 50287459, "url": "https://news.ycombinator.com/item?id=50287459", "by": "jeremyjh", "date": "2026-10-29", "stance": "recommend", "quote": "If you like to customize your tools Pi is the best", "excerpt": "Every harness has skills & MCPs. I think omp.sh (oh-my-pi) is the best by a comfortable margin and its completely batteries included. It will pick up and use your Claude Code skills and plugins by default. If you like to customize your tools Pi is the best, it has a huge plugin ecosystem (all can be used in OMP as well, but it has opinionated answers for most things already). Stay away from OpenCode, its a pile of sl", "context": "", "use_case": "customizable AI coding harness", "switched_from": "", "story": "Claude Code regularly sends each background agent\u2019s transcript to its model"}]}, {"text": ", plus token-efficient context handling that works with both local models and Claude or GPT subscriptions.", "cites": [{"id": 798842672, "item_id": 49927667, "url": "https://news.ycombinator.com/item?id=49927667", "by": "darklinear", "date": "2026-10-01", "stance": "recommend", "quote": "I find that it does better than the native codex harness", "excerpt": "I use pi with a ChatGPT subscription. I find that it does better than the native codex harness: - it's very token efficient. I've had whole profiling + refactoring sessions finish in <20k tokens, whereas in codex you can get close to that with just all the cruft OpenAI loads by default before you even send your own messages - the token efficiency means turns are a lot snappier - fine-grained context control by defaul", "context": "", "use_case": "profiling and refactoring with ChatGPT subscription", "switched_from": "Codex", "story": "Pi 1.0"}, {"id": 798924704, "item_id": 49932794, "url": "https://news.ycombinator.com/item?id=49932794", "by": "nfriedly", "date": "2026-10-02", "stance": "recommend", "quote": "I really like Pi, especially for when you're running a local model and your context is particularly limited.", "excerpt": "I really like Pi, especially for when you're running a local model and your context is particularly limited. But even when running a big model, I think building up your environment to be exactly what you want can be really powerful. When I was first getting started and didn't realize that ollama defaults to incredibly tiny amounts of context, Pi was the only harness that actually worked because the rest of them used ", "context": "", "use_case": "coding harness for local models", "switched_from": "", "story": "Pi 1.0"}, {"id": 804487872, "item_id": 50280492, "url": "https://news.ycombinator.com/item?id=50280492", "by": "duckqlz", "date": "2026-10-28", "stance": "recommend", "quote": "Pi can utilize Claude or GPT subscriptions and they work wonderfully.", "excerpt": "Pi can utilize Claude or GPT subscriptions and they work wonderfully. You can have both subscriptions logged in and write your own tools to route to the model that works best in every scenario. I have access to both pro level subscriptions for work and a large credit count for api calls for Claude. I use all 3 with JEV as my router with a local model for context management and compaction. my token usage has gone down", "context": "has work subscriptions to Claude and GPT", "use_case": "open-source coding harness", "switched_from": "", "story": "The Agent SDK"}, {"id": 798868401, "item_id": 49929275, "url": "https://news.ycombinator.com/item?id=49929275", "by": "pettijohn", "date": "2026-10-02", "stance": "recommend", "quote": "It's been so easy to bend into what I need.", "excerpt": "I use PI WEB (https://pi-web.dev, there's more than one with that name) to orchestrate remotely. Pi lets me run local models (currently Qwen 3.8 Flash Next on Strix Halo 128GB) and flagship models (with subscription auth, not API pricing) side by side. I typically ask GPT 6.1 Sol to review requirements and then spawn a subsession with Qwen, review Qwen's work, ask Qwen to fix. About 95% success with a single pass lik", "context": "runs local models on Strix Halo 128GB", "use_case": "mixing local and flagship models", "switched_from": "", "story": "Pi 1.0"}]}, {"text": " The most common complaint is a messy plugin ecosystem", "cites": [{"id": 802691729, "item_id": 50168233, "url": "https://news.ycombinator.com/item?id=50168233", "by": "verdverm", "date": "2026-10-20", "stance": "warn", "quote": "Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries.", "excerpt": "Why are plugins a desirable feature for core features? Today I use GoModel [1], and in example, all providers are built in. Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries. Can you explain why you think a plugin based system is preferable to a competent team building a complete, curated product? Plugins seems like a great way to expose yourself to supply chain security ", "context": "", "use_case": "", "switched_from": "", "story": "Launch HN: Respan (YC W24) \u2013 Open-source Rust AI gateway built around plugins"}, {"id": 798922816, "item_id": 49932676, "url": "https://news.ycombinator.com/item?id=49932676", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-coding projects than seriously maintained tools.", "excerpt": "Long-time pi user here. If you're thinking of trying out pi but feel daunted by finding the right plugins to get started, I wrote about my minimal pi setup here: https://github.com/richardgill/pi-extensions/blob/main/PI_SETUP.md It's still pretty minimal and brings back just enough stuff that I don't miss Claude Code. Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-codi", "context": "Long-time pi user", "use_case": "AI coding agent", "switched_from": "", "story": "Pi 1.0"}]}, {"text": ", where many extensions feel vibe-coded or bloated and raise security worries.", "cites": [{"id": 798923888, "item_id": 49932743, "url": "https://news.ycombinator.com/item?id=49932743", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "I've daily driven Pi for 9+ months and I recommend pi a lot to friends.", "excerpt": "Posted this elsewhere in this thread. I've daily driven Pi for 9+ months and I recommend pi a lot to friends. I think switching to pi from claude code is pretty daunting. The pi extension system is excellent. But which extensions should you use? And honestly a lot of the most popular extensions to me feel bloated / weird / vibe-coded. Here's my minimal set up: https://github.com/richardgill/pi-extensions/blob/main/PI", "context": "daily driver for 9+ months", "use_case": "coding agent", "switched_from": "Claude Code", "story": "Pi 1.0"}, {"id": 798922816, "item_id": 49932676, "url": "https://news.ycombinator.com/item?id=49932676", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-coding projects than seriously maintained tools.", "excerpt": "Long-time pi user here. If you're thinking of trying out pi but feel daunted by finding the right plugins to get started, I wrote about my minimal pi setup here: https://github.com/richardgill/pi-extensions/blob/main/PI_SETUP.md It's still pretty minimal and brings back just enough stuff that I don't miss Claude Code. Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-codi", "context": "Long-time pi user", "use_case": "AI coding agent", "switched_from": "", "story": "Pi 1.0"}, {"id": 802691729, "item_id": 50168233, "url": "https://news.ycombinator.com/item?id=50168233", "by": "verdverm", "date": "2026-10-20", "stance": "warn", "quote": "Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries.", "excerpt": "Why are plugins a desirable feature for core features? Today I use GoModel [1], and in example, all providers are built in. Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries. Can you explain why you think a plugin based system is preferable to a competent team building a complete, curated product? Plugins seems like a great way to expose yourself to supply chain security ", "context": "", "use_case": "", "switched_from": "", "story": "Launch HN: Respan (YC W24) \u2013 Open-source Rust AI gateway built around plugins"}]}], "praise": [{"label": "Extensibility", "summary": "Easy to customize, and Pi can write its own plugins because it knows its own docs.", "quotes": [{"id": 798898528, "item_id": 49931158, "url": "https://news.ycombinator.com/item?id=49931158", "by": "eloisant", "date": "2026-10-02", "stance": "recommend", "quote": "I switched from OpenCode to Pi and never looked back.", "excerpt": "Yes, I switched from OpenCode to Pi and never looked back. What's great with Pi is that it's very easy to extend, because it knows its own doc. So you tell it \"implement a plugin that does this\" and it does it directly in its own folder. That means I have a plan mode that works exactly the way I want, I have a hook that cleans up added comments after each change, I can ask \"pull this github PR and assess the comments", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "Pi 1.0"}, {"id": 804599345, "item_id": 50287459, "url": "https://news.ycombinator.com/item?id=50287459", "by": "jeremyjh", "date": "2026-10-29", "stance": "recommend", "quote": "If you like to customize your tools Pi is the best", "excerpt": "Every harness has skills & MCPs. I think omp.sh (oh-my-pi) is the best by a comfortable margin and its completely batteries included. It will pick up and use your Claude Code skills and plugins by default. If you like to customize your tools Pi is the best, it has a huge plugin ecosystem (all can be used in OMP as well, but it has opinionated answers for most things already). Stay away from OpenCode, its a pile of sl", "context": "", "use_case": "customizable AI coding harness", "switched_from": "", "story": "Claude Code regularly sends each background agent\u2019s transcript to its model"}, {"id": 799595216, "item_id": 49974701, "url": "https://news.ycombinator.com/item?id=49974701", "by": "clx75", "date": "2026-10-06", "stance": "recommend", "quote": "I use Pi and created an extension which teaches the LLM how to start a JVM with a Clojure nREPL", "excerpt": "I have similar experiences in Clojure. I use Pi and created an extension which teaches the LLM how to start a JVM with a Clojure nREPL listening in it. Added a tool which lets the agent evaluate any Clojure form inside the JVM. Sprinkled the setup with clj-reload and now the agent is using the nREPL as if it were its second nature. What I like most is how fast testing goes: it modifies test code on disk, asks clj-rel", "context": "Clojure developer", "use_case": "Clojure agent development", "switched_from": "", "story": "Why Common Lisp is now the best programming language"}]}, {"label": "Lightweight minimalism", "summary": "Uses far less RAM and CPU than OpenCode and feels light and nimble.", "quotes": [{"id": 801584176, "item_id": 50099011, "url": "https://news.ycombinator.com/item?id=50099011", "by": "usernametaken29", "date": "2026-10-15", "stance": "recommend", "quote": "I recently switched to pi which uses an order of magnitude less RAM (hundreds vs dozens)", "excerpt": "I wish they\u2019d invest more time in performance improvements over code mods. I recently switched to pi which uses an order of magnitude less RAM (hundreds vs dozens) and noticeably less CPU on my machine\u2026 OpenCode is really nice but it\u2019s about as heavy as antigravity and is substandard in features\u2026", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 798849136, "item_id": 49928071, "url": "https://news.ycombinator.com/item?id=49928071", "by": "lo5", "date": "2026-10-01", "stance": "recommend", "quote": "It feels light and nimble.", "excerpt": "I use Pi with only a couple of extensions and using less and less of CC and Codex everyday. It feels light and nimble. Codemode from yesterday's release made it even better. I wish I could use my Claude subscription via Pi.", "context": "", "use_case": "AI coding agent", "switched_from": "Claude Code", "story": "Pi 1.0"}, {"id": 799654720, "item_id": 49978420, "url": "https://news.ycombinator.com/item?id=49978420", "by": "Roark66", "date": "2026-10-06", "stance": "recommend", "quote": "With open weight models I use Pi coding agent with a self written auto permission and Web search plugins.", "excerpt": "With open weight models I use Pi coding agent with a self written auto permission and Web search plugins. I'm happy with it, but the plugins are very important. I used Open Code long time ago, but I prefer the minimalism of pi.", "context": "", "use_case": "coding agent with open weight models", "switched_from": "OpenCode", "story": "Several vulnerabilities have been discovered in the Linux kernel"}]}, {"label": "Model flexibility", "summary": "Runs local models and Claude or GPT subscriptions side by side, without lock-in to one provider.", "quotes": [{"id": 804487872, "item_id": 50280492, "url": "https://news.ycombinator.com/item?id=50280492", "by": "duckqlz", "date": "2026-10-28", "stance": "recommend", "quote": "Pi can utilize Claude or GPT subscriptions and they work wonderfully.", "excerpt": "Pi can utilize Claude or GPT subscriptions and they work wonderfully. You can have both subscriptions logged in and write your own tools to route to the model that works best in every scenario. I have access to both pro level subscriptions for work and a large credit count for api calls for Claude. I use all 3 with JEV as my router with a local model for context management and compaction. my token usage has gone down", "context": "has work subscriptions to Claude and GPT", "use_case": "open-source coding harness", "switched_from": "", "story": "The Agent SDK"}, {"id": 798868401, "item_id": 49929275, "url": "https://news.ycombinator.com/item?id=49929275", "by": "pettijohn", "date": "2026-10-02", "stance": "recommend", "quote": "It's been so easy to bend into what I need.", "excerpt": "I use PI WEB (https://pi-web.dev, there's more than one with that name) to orchestrate remotely. Pi lets me run local models (currently Qwen 3.8 Flash Next on Strix Halo 128GB) and flagship models (with subscription auth, not API pricing) side by side. I typically ask GPT 6.1 Sol to review requirements and then spawn a subsession with Qwen, review Qwen's work, ask Qwen to fix. About 95% success with a single pass lik", "context": "runs local models on Strix Halo 128GB", "use_case": "mixing local and flagship models", "switched_from": "", "story": "Pi 1.0"}, {"id": 798901280, "item_id": 49931330, "url": "https://news.ycombinator.com/item?id=49931330", "by": "dozerly", "date": "2026-10-02", "stance": "recommend", "quote": "I use Pi because I fundamentally believe that inference providers are not the correct providers of harnesses.", "excerpt": "I use Pi because I fundamentally believe that inference providers are not the correct providers of harnesses. Their incentives are not aligned, and inevitably Claude and codex will continue trying to lock you in to get your inference spend.", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}]}, {"label": "Token efficiency", "summary": "A small system prompt and fine-grained context control save tokens, which helps with limited local contexts.", "quotes": [{"id": 798842672, "item_id": 49927667, "url": "https://news.ycombinator.com/item?id=49927667", "by": "darklinear", "date": "2026-10-01", "stance": "recommend", "quote": "I find that it does better than the native codex harness", "excerpt": "I use pi with a ChatGPT subscription. I find that it does better than the native codex harness: - it's very token efficient. I've had whole profiling + refactoring sessions finish in <20k tokens, whereas in codex you can get close to that with just all the cruft OpenAI loads by default before you even send your own messages - the token efficiency means turns are a lot snappier - fine-grained context control by defaul", "context": "", "use_case": "profiling and refactoring with ChatGPT subscription", "switched_from": "Codex", "story": "Pi 1.0"}, {"id": 798924704, "item_id": 49932794, "url": "https://news.ycombinator.com/item?id=49932794", "by": "nfriedly", "date": "2026-10-02", "stance": "recommend", "quote": "I really like Pi, especially for when you're running a local model and your context is particularly limited.", "excerpt": "I really like Pi, especially for when you're running a local model and your context is particularly limited. But even when running a big model, I think building up your environment to be exactly what you want can be really powerful. When I was first getting started and didn't realize that ollama defaults to incredibly tiny amounts of context, Pi was the only harness that actually worked because the rest of them used ", "context": "", "use_case": "coding harness for local models", "switched_from": "", "story": "Pi 1.0"}, {"id": 803654833, "item_id": 50228427, "url": "https://news.ycombinator.com/item?id=50228427", "by": "badsectoracula", "date": "2026-10-25", "stance": "recommend", "quote": "Pi's \"rolling\" approach to compaction (i.e. summarizing the older messages to make space while keeping the recent messages as-is) make it almost a non-issue.", "excerpt": "Interesting but (from the PDF): Every role but the verifier is the same underlying model, called in a fresh context and coordinating through a shared workspace on disk: <ws>/task.md - the problem statement <ws>/plan.md - the manager\u2019s overarching plan <ws>/tasks.json - a debug mirror of the task list <ws>/notes.md - curated ideas / findings / partial proofs <ws>/solution.py - the current solution Then they describe h", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Five Qwen3.8-27B models match Fable 5 Using GVS5H Harness"}]}, {"label": "Stability", "summary": "Frequent updates rarely break anything or change the UX.", "quotes": [{"id": 801011936, "item_id": 50063246, "url": "https://news.ycombinator.com/item?id=50063246", "by": "spudlyo", "date": "2026-10-12", "stance": "recommend", "quote": "I've been using Pi nearly every day for 4-5 months now", "excerpt": "I've been using Pi nearly every day for 4-5 months now, and even though it is still very much in development with new updates arriving nearly every week, it has to my mind never regressed. I've never lamented that a Pi update ever broke anything I depend on or made anything worse. I immediately adopted the new tuiMode \"fullscreen\" feature the moment it landed, and it's been exceptional. I love the way it now manages ", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 798822848, "item_id": 49926428, "url": "https://news.ycombinator.com/item?id=49926428", "by": "lionkor", "date": "2026-10-01", "stance": "recommend", "quote": "I love how consistent pi has been, especially in regards to not breaking ux", "excerpt": "I love how consistent pi has been, especially in regards to not breaking ux", "context": "", "use_case": "", "switched_from": "", "story": "Pi 1.0"}, {"id": 798822352, "item_id": 49926397, "url": "https://news.ycombinator.com/item?id=49926397", "by": "sroerick", "date": "2026-10-01", "stance": "recommend", "quote": "Pi is really good. I use it for a majority of my work.", "excerpt": "Pi is really good. I use it for a majority of my work. I also like Autolith. The freedom of having a lisp machine is, to me, much more enjoyable than trying to maintain Typescript. But I'm not a typescript guy. I do largely stick with Pi because it's very bulletproofed", "context": "", "use_case": "", "switched_from": "", "story": "Pi 1.0"}]}, {"label": "Transparency", "summary": "Shows context use and cost, and keeps its config in inspectable folders on disk.", "quotes": [{"id": 798878752, "item_id": 49929922, "url": "https://news.ycombinator.com/item?id=49929922", "by": "hermannj314", "date": "2026-10-02", "stance": "recommend", "quote": "it was the joy a using a harness that was transparent about how it worked.", "excerpt": "I am using pi the same way everyone uses Codex and Claude, but pi makes it very simple to make folders to make the agent behave a certain way and that was easy for me to understand. ChatGPT Desktop seems to intentionally hide what they are doing and where they are storing state, sessions, config, I never had any idea what is in context or what \"archiving\" even means. I got sick of OpenAI hiding how ChatGPT Desktop wa", "context": "", "use_case": "AI coding harness", "switched_from": "ChatGPT Desktop", "story": "Pi 1.0"}, {"id": 798902304, "item_id": 49931394, "url": "https://news.ycombinator.com/item?id=49931394", "by": "ttmacer", "date": "2026-10-02", "stance": "recommend", "quote": "I kinda like having this abstraction on my filesystem to inspect instead of it being hidden away in a opaque harness", "excerpt": "AFAIK Pi out of the box as other harnesses handles this use case via dot directories on your filesystem. So a .pi directory in your working directory Pi gets started in is the place where you would implement such a profile. Note that there is also a global .pi in the home directory, which takes precedence over session-local .pi directories. In the past I used this abstraction conveniently to setup different agents fo", "context": "", "use_case": "configurable AI agent profiles", "switched_from": "", "story": "Pi 1.0"}]}], "caveats": [{"label": "Plugin quality", "summary": "Many plugins feel like unmaintained personal vibe-coding projects, and some users have security worries.", "quotes": [{"id": 798922816, "item_id": 49932676, "url": "https://news.ycombinator.com/item?id=49932676", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-coding projects than seriously maintained tools.", "excerpt": "Long-time pi user here. If you're thinking of trying out pi but feel daunted by finding the right plugins to get started, I wrote about my minimal pi setup here: https://github.com/richardgill/pi-extensions/blob/main/PI_SETUP.md It's still pretty minimal and brings back just enough stuff that I don't miss Claude Code. Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-codi", "context": "Long-time pi user", "use_case": "AI coding agent", "switched_from": "", "story": "Pi 1.0"}, {"id": 802691729, "item_id": 50168233, "url": "https://news.ycombinator.com/item?id=50168233", "by": "verdverm", "date": "2026-10-20", "stance": "warn", "quote": "Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries.", "excerpt": "Why are plugins a desirable feature for core features? Today I use GoModel [1], and in example, all providers are built in. Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries. Can you explain why you think a plugin based system is preferable to a competent team building a complete, curated product? Plugins seems like a great way to expose yourself to supply chain security ", "context": "", "use_case": "", "switched_from": "", "story": "Launch HN: Respan (YC W24) \u2013 Open-source Rust AI gateway built around plugins"}, {"id": 798923888, "item_id": 49932743, "url": "https://news.ycombinator.com/item?id=49932743", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "I've daily driven Pi for 9+ months and I recommend pi a lot to friends.", "excerpt": "Posted this elsewhere in this thread. I've daily driven Pi for 9+ months and I recommend pi a lot to friends. I think switching to pi from claude code is pretty daunting. The pi extension system is excellent. But which extensions should you use? And honestly a lot of the most popular extensions to me feel bloated / weird / vibe-coded. Here's my minimal set up: https://github.com/richardgill/pi-extensions/blob/main/PI", "context": "daily driver for 9+ months", "use_case": "coding agent", "switched_from": "Claude Code", "story": "Pi 1.0"}]}, {"label": "Daunting setup", "summary": "Switching from Claude Code is daunting because it is hard to know which extensions to use.", "quotes": [{"id": 798923888, "item_id": 49932743, "url": "https://news.ycombinator.com/item?id=49932743", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "I've daily driven Pi for 9+ months and I recommend pi a lot to friends.", "excerpt": "Posted this elsewhere in this thread. I've daily driven Pi for 9+ months and I recommend pi a lot to friends. I think switching to pi from claude code is pretty daunting. The pi extension system is excellent. But which extensions should you use? And honestly a lot of the most popular extensions to me feel bloated / weird / vibe-coded. Here's my minimal set up: https://github.com/richardgill/pi-extensions/blob/main/PI", "context": "daily driver for 9+ months", "use_case": "coding agent", "switched_from": "Claude Code", "story": "Pi 1.0"}, {"id": 798922816, "item_id": 49932676, "url": "https://news.ycombinator.com/item?id=49932676", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-coding projects than seriously maintained tools.", "excerpt": "Long-time pi user here. If you're thinking of trying out pi but feel daunted by finding the right plugins to get started, I wrote about my minimal pi setup here: https://github.com/richardgill/pi-extensions/blob/main/PI_SETUP.md It's still pretty minimal and brings back just enough stuff that I don't miss Claude Code. Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-codi", "context": "Long-time pi user", "use_case": "AI coding agent", "switched_from": "", "story": "Pi 1.0"}]}, {"label": "Creeping bloat", "summary": "Some users find it bulkier than hoped and doubt its minimalism claims as features accumulate.", "quotes": [{"id": 804937153, "item_id": 50308572, "url": "https://news.ycombinator.com/item?id=50308572", "by": "ravenstine", "date": "2026-10-30", "stance": "warn", "quote": "even that ended up being more of a bulky mess than what I was hoping.", "excerpt": "Not currently, but I probably will be in the coming weeks. I pretty much end up hating every harness I use. Some come really close to what I'm looking for but then completely miss the mark in other ways. At the moment, I'm stripping OpenCode down to its absolute bare minimum, removing pretty much all of the vendor-specific crap and everything else I don't plan on using. Once I'm convinced that's stable, I'll identify", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Ask HN: Are you using a custom harness?"}, {"id": 798862320, "item_id": 49928895, "url": "https://news.ycombinator.com/item?id=49928895", "by": "zmmmmm", "date": "2026-10-02", "stance": "recommend", "quote": "I love Pi", "excerpt": "I love Pi but I am sceptical of their claim to minimalism. New tools often make such claims as an excuse for not having a lot of features. You didn't want those features anyway! As they mature, the features and complexity creep in and before you know it, the pitch changes to more of a full stack one. I don't mind though, because I think either way it leads to a better design under the hood when things are built to be", "context": "", "use_case": "", "switched_from": "", "story": "Pi 1.0"}]}, {"label": "Fullscreen default", "summary": "One user objected to the new full-screen mode replacing native terminal scrolling by default.", "quotes": [{"id": 798872784, "item_id": 49929549, "url": "https://news.ycombinator.com/item?id=49929549", "by": "octoberfranklin", "date": "2026-10-02", "stance": "warn", "quote": "The main reason I use Pi is because it didn't try to reinvent my terminal's buttery-smooth native scrolling", "excerpt": "Full-screen mode by default Man, first MCP then this? The main reason I use Pi is because it didn't try to reinvent my terminal's buttery-smooth native scrolling -- something no terminal client can ever match. Did you all get acquired by private equity or something? The enshittification is coming at us fast.", "context": "", "use_case": "terminal coding agent", "switched_from": "", "story": "Pi 1.0"}]}], "who": [{"by": "duckqlz", "context": "has work subscriptions to Claude and GPT", "url": "https://news.ycombinator.com/item?id=50280492", "date": "2026-10-28"}, {"by": "nrclark", "context": "software developer", "url": "https://news.ycombinator.com/item?id=50261642", "date": "2026-10-27"}, {"by": "paid_dot_expert", "context": "on $100/m Claude sub", "url": "https://news.ycombinator.com/item?id=50151055", "date": "2026-10-19"}, {"by": "aizk", "context": "works for an open source AI inference provider", "url": "https://news.ycombinator.com/item?id=50135410", "date": "2026-10-17"}, {"by": "sanex", "context": "former engineering manager, now IC", "url": "https://news.ycombinator.com/item?id=50130991", "date": "2026-10-17"}, {"by": "verdverm", "context": "former Vim user now on VS Code", "url": "https://news.ycombinator.com/item?id=50097171", "date": "2026-10-14"}, {"by": "JLO64", "context": "prefers minimizing extensions", "url": "https://news.ycombinator.com/item?id=50013187", "date": "2026-10-08"}, {"by": "clx75", "context": "Clojure developer", "url": "https://news.ycombinator.com/item?id=49974701", "date": "2026-10-06"}, {"by": "Schlagbohrer", "context": "uses a monitor tipped 90\u00b0", "url": "https://news.ycombinator.com/item?id=49961324", "date": "2026-10-05"}, {"by": "theturtletalks", "context": "built a software factory", "url": "https://news.ycombinator.com/item?id=49939357", "date": "2026-10-02"}, {"by": "richardgill88", "context": "daily driver for 9+ months", "url": "https://news.ycombinator.com/item?id=49932743", "date": "2026-10-02"}, {"by": "alexfortin", "context": "maintains Pi extensions", "url": "https://news.ycombinator.com/item?id=49931485", "date": "2026-10-02"}], "tags": ["for local models", "for tinkerers", "for terminal users", "for multi-model setups", "open source"], "use_cases": ["coding agent harness", "coding agent", "agent harness", "ai coding agent", "ai coding harness", "coding harness"], "compared": [{"slug": "claude-code", "name": "Claude Code", "co": 123, "switched_from": 23, "switched_to": 1}, {"slug": "opencode", "name": "OpenCode", "co": 112, "switched_from": 23, "switched_to": 3}, {"slug": "codex", "name": "Codex", "co": 80, "switched_from": 8, "switched_to": 0}, {"slug": "oh-my-pi", "name": "oh-my-pi", "co": 30, "switched_from": 3, "switched_to": 5}, {"slug": "claude", "name": "Claude", "co": 33, "switched_from": 4, "switched_to": 0}, {"slug": "deepseek", "name": "DeepSeek", "co": 35, "switched_from": 0, "switched_to": 0}, {"slug": "claude-opus", "name": "Claude Opus", "co": 27, "switched_from": 0, "switched_to": 0}, {"slug": "deepseek-v4-flash", "name": "DeepSeek V4 Flash", "co": 25, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=49926069", "title": "Pi 1.0", "mentions": 42}, {"url": "https://news.ycombinator.com/item?id=49176038", "title": "Pi's Minimalism Is Its Advantage", "mentions": 22}, {"url": "https://news.ycombinator.com/item?id=50094994", "title": "OpenCode 2.0", "mentions": 12}, {"url": "https://news.ycombinator.com/item?id=48542100", "title": "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?", "mentions": 10}, {"url": "https://news.ycombinator.com/item?id=48883275", "title": "Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k", "mentions": 9}, {"url": "https://news.ycombinator.com/item?id=49906637", "title": "You said no MCP", "mentions": 7}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {"recommend": 1}, "2026-02": {"compare": 1, "recommend": 11}, "2026-03": {"recommend": 1, "neutral": 2}, "2026-04": {"recommend": 12, "compare": 1, "neutral": 4}, "2026-05": {"recommend": 5, "neutral": 7}, "2026-06": {"recommend": 42, "neutral": 25, "compare": 2, "warn": 2}, "2026-07": {"recommend": 42, "compare": 4, "neutral": 27, "warn": 1}, "2026-08": {"warn": 2, "recommend": 48, "neutral": 23, "compare": 8}, "2026-09": {"recommend": 30, "neutral": 23, "compare": 2, "warn": 2}, "2026-10": {"neutral": 25, "recommend": 53, "compare": 8, "warn": 2}}, "quotes": [{"id": 805051377, "item_id": 50315711, "url": "https://news.ycombinator.com/item?id=50315711", "by": "paper2d", "date": "2026-10-31", "stance": "recommend", "quote": "I used deepseek and pi harness", "excerpt": "I did something similar with an rare, old 32bit game I really enjoyed in my childhood. I used deepseek and pi harness and it decompiled it via ghidra and recompiled it for wasm in under 40 mins or so and made it run in the browser. I was impressed! For anyone who wants to run the game in your browser, here it is - https://bonkheads-port-by-kapil.pages.dev/", "context": "", "use_case": "agent harness", "switched_from": "", "story": "Cisco gives engineer unlimited AI tokens to rewrite legacy code"}, {"id": 804937153, "item_id": 50308572, "url": "https://news.ycombinator.com/item?id=50308572", "by": "ravenstine", "date": "2026-10-30", "stance": "warn", "quote": "even that ended up being more of a bulky mess than what I was hoping.", "excerpt": "Not currently, but I probably will be in the coming weeks. I pretty much end up hating every harness I use. Some come really close to what I'm looking for but then completely miss the mark in other ways. At the moment, I'm stripping OpenCode down to its absolute bare minimum, removing pretty much all of the vendor-specific crap and everything else I don't plan on using. Once I'm convinced that's stable, I'll identify", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Ask HN: Are you using a custom harness?"}, {"id": 804906769, "item_id": 50306673, "url": "https://news.ycombinator.com/item?id=50306673", "by": "kyle-avrett", "date": "2026-10-30", "stance": "neutral", "quote": "ricing herdr-app with pi", "excerpt": "Just finished ricing herdr-app with pi, damn it looks good! https://github.com/kyle-avrett/herdr-square-dracula", "context": "", "use_case": "customizing app theme", "switched_from": "", "story": "Herdr has released native app"}, {"id": 804599345, "item_id": 50287459, "url": "https://news.ycombinator.com/item?id=50287459", "by": "jeremyjh", "date": "2026-10-29", "stance": "recommend", "quote": "If you like to customize your tools Pi is the best", "excerpt": "Every harness has skills & MCPs. I think omp.sh (oh-my-pi) is the best by a comfortable margin and its completely batteries included. It will pick up and use your Claude Code skills and plugins by default. If you like to customize your tools Pi is the best, it has a huge plugin ecosystem (all can be used in OMP as well, but it has opinionated answers for most things already). Stay away from OpenCode, its a pile of sl", "context": "", "use_case": "customizable AI coding harness", "switched_from": "", "story": "Claude Code regularly sends each background agent\u2019s transcript to its model"}, {"id": 804487872, "item_id": 50280492, "url": "https://news.ycombinator.com/item?id=50280492", "by": "duckqlz", "date": "2026-10-28", "stance": "recommend", "quote": "Pi can utilize Claude or GPT subscriptions and they work wonderfully.", "excerpt": "Pi can utilize Claude or GPT subscriptions and they work wonderfully. You can have both subscriptions logged in and write your own tools to route to the model that works best in every scenario. I have access to both pro level subscriptions for work and a large credit count for api calls for Claude. I use all 3 with JEV as my router with a local model for context management and compaction. my token usage has gone down", "context": "has work subscriptions to Claude and GPT", "use_case": "open-source coding harness", "switched_from": "", "story": "The Agent SDK"}, {"id": 804423792, "item_id": 50276487, "url": "https://news.ycombinator.com/item?id=50276487", "by": "krzs9", "date": "2026-10-28", "stance": "neutral", "quote": "I use Pi as my harness and they run in herdr.", "excerpt": "I have a Hetzner box that costs about 10$ a month - I use that as my always-on box. I use Pi as my harness and they run in herdr. I can then from my Mac (bottom of the line macbook air - next laptop will be a Neo) ssh to the box and control my agents. I can also ssh into the box from my iPhone (using Termius) and continue seamlessly from there too.", "context": "", "use_case": "agent harness", "switched_from": "", "story": "New 14\u2011inch MacBook Pro with M6"}, {"id": 804186273, "item_id": 50261642, "url": "https://news.ycombinator.com/item?id=50261642", "by": "nrclark", "date": "2026-10-27", "stance": "recommend", "quote": "I like Pi, personally", "excerpt": "That might be true for general consumers, yes. Not so true for software developers though. For developers, the tooling is definitely not a moat, and that means the models are what must differentiate. Claude code is very polished, but I can get the same jobs done using Cursor CLI or any number of open-source coding harnesses (I like Pi, personally). Most plugins/skills/mcps are portable between harnesses with minimal ", "context": "software developer", "use_case": "open-source coding harness", "switched_from": "", "story": "Anthropic S-1"}, {"id": 804099106, "item_id": 50256194, "url": "https://news.ycombinator.com/item?id=50256194", "by": "Degorath", "date": "2026-10-27", "stance": "compare", "quote": "of the set {Claude Code, Codex, Pi, OMP} I must say I find CC to be the worst.", "excerpt": "That's quite surprising, of the set {Claude Code, Codex, Pi, OMP} I must say I find CC to be the worst. They all have some annoying downsides and I use CC the most, but it's gotten to the point where I'm for sure actively shopping around for better harnesses.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Anthropic S-1"}, {"id": 803856545, "item_id": 50241034, "url": "https://news.ycombinator.com/item?id=50241034", "by": "crossroadsguy", "date": "2026-10-26", "stance": "compare", "quote": "At least Pi is completely stripped (it might be changing though) and then you can make it as you want it.", "excerpt": "At this point I'd rather try to invest effort and usage in a harness that is FOSS and is actually steered by a non-profit. Goose comes to mind and to be honest I haven't tried it beyond installing and giving it a prompt or two. Comiong to harnesses, even Pi seems like a weird choice. I mean Pi, until it's \"your own customised Pi\", is pretty useless and pointless. Besides they are also starting to have their own model", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?"}, {"id": 803654833, "item_id": 50228427, "url": "https://news.ycombinator.com/item?id=50228427", "by": "badsectoracula", "date": "2026-10-25", "stance": "recommend", "quote": "Pi's \"rolling\" approach to compaction (i.e. summarizing the older messages to make space while keeping the recent messages as-is) make it almost a non-issue.", "excerpt": "Interesting but (from the PDF): Every role but the verifier is the same underlying model, called in a fresh context and coordinating through a shared workspace on disk: <ws>/task.md - the problem statement <ws>/plan.md - the manager\u2019s overarching plan <ws>/tasks.json - a debug mirror of the task list <ws>/notes.md - curated ideas / findings / partial proofs <ws>/solution.py - the current solution Then they describe h", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Five Qwen3.8-27B models match Fable 5 Using GVS5H Harness"}, {"id": 803538657, "item_id": 50221166, "url": "https://news.ycombinator.com/item?id=50221166", "by": "kelvinjps10", "date": "2026-10-24", "stance": "recommend", "quote": "20$ chatgpt plus sub with pi harness", "excerpt": "20$ chatgpt plus sub with pi harness and a few api credits in openrouter.sometiemd I use sol for planing and glm flash for implementation bu", "context": "", "use_case": "agent harness", "switched_from": "", "story": "Kimi K3.1 Pro and K3.1 Pro Highspeed"}, {"id": 803487921, "item_id": 50217995, "url": "https://news.ycombinator.com/item?id=50217995", "by": "eleventen", "date": "2026-10-24", "stance": "compare", "quote": "Herdr\u2019s plugin system is quite limited and immature compared to something like Pi.", "excerpt": "Probably very little. Herdr\u2019s plugin system is quite limited and immature compared to something like Pi. I love herdr and use it daily but it has its limits. Nice to see people experimenting. I\u2019m not really an early adopter but hopefully the good ideas get picked up by herdr", "context": "", "use_case": "plugin system", "switched_from": "", "story": "Show HN: Corgi, a helpful {agent, terminal} multiplexer"}, {"id": 803194353, "item_id": 50199647, "url": "https://news.ycombinator.com/item?id=50199647", "by": "thehours", "date": "2026-10-22", "stance": "neutral", "quote": "make me comfortable running Pi in YOLO mode", "excerpt": "In this vein, Anthropic's sandbox-runtime [0] has struck a good balance between security and convenience. Simple config, filesystem and network restrictions, and credential masking make me comfortable running Pi in YOLO mode. Though I'm often running into frustrating (albeit intentional) sandbox restrictions with common workflows e.g. xcode/swift and Docker access. Would love suggestions for how to give agents access", "context": "", "use_case": "coding agent", "switched_from": "", "story": "OpenShell \u2013 A safe, private runtime for fleets of autonomous AI agents"}, {"id": 802802737, "item_id": 50175171, "url": "https://news.ycombinator.com/item?id=50175171", "by": "0xbadcafebee", "date": "2026-10-21", "stance": "neutral", "quote": "we need 1,000 different Pi extensions that will intervene as the agent codes", "excerpt": "> The full benchmark is hard in a way that matters to me: the agents get the main behavior right much more often than they keep all tested behavior working as requirements change So, it's like... software development? 90% of bugs come from unintended consequences of changes after the fact. There's a couple decades of design lessons people had to learn to keep this number low. The agents don't follow those lessons. Go", "context": "", "use_case": "enforcing design in coding agents", "switched_from": "", "story": "SlopCodeBench 10/2026 \u2013 Opus 5.5, Astra wins overall, xhigh considered harmful"}, {"id": 802691729, "item_id": 50168233, "url": "https://news.ycombinator.com/item?id=50168233", "by": "verdverm", "date": "2026-10-20", "stance": "warn", "quote": "Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries.", "excerpt": "Why are plugins a desirable feature for core features? Today I use GoModel [1], and in example, all providers are built in. Having recently given Pi a try, the plugin ecosystem is a mess and I have more security worries. Can you explain why you think a plugin based system is preferable to a competent team building a complete, curated product? Plugins seems like a great way to expose yourself to supply chain security ", "context": "", "use_case": "", "switched_from": "", "story": "Launch HN: Respan (YC W24) \u2013 Open-source Rust AI gateway built around plugins"}, {"id": 802418083, "item_id": 50151130, "url": "https://news.ycombinator.com/item?id=50151130", "by": "CharlieDigital", "date": "2026-10-19", "stance": "neutral", "quote": "Other harnesses like OpenCode, Pi, etc. you can hook into their plugin architecture and implement your own OTEL emission", "excerpt": "You don't need a proxy to do this. I would argue you should not use a proxy to do this because the end result is a single point of failure direct in the line of your productivity. Both Codex[0] and Claude Code[1] output OTEL traces that you can ingest and process. Other harnesses like OpenCode, Pi, etc. you can hook into their plugin architecture and implement your own OTEL emission from a plugin (or use an off the s", "context": "", "use_case": "OTEL emission via plugin", "switched_from": "", "story": "If you don't run your own LLM proxy, how do you know whats going on?"}, {"id": 802416880, "item_id": 50151055, "url": "https://news.ycombinator.com/item?id=50151055", "by": "paid_dot_expert", "date": "2026-10-19", "stance": "neutral", "quote": "Asked Pi what I used last 3 days.", "excerpt": "Asked Pi what I used last 3 days. \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510 \u2502 UTC day \u2502 Cost \u2502 \u251c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2524 \u2502 Oct 16 \u2502 $2.72 \u2502 \u251c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2524 \u2502 Oct 17 \u2502 $300.82 \u2502 \u251c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2524 \u2502 Oct 18 \u2502 $208.72 \u2502 \u251c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2524 \u2502 Oct 19 (so far) \u2502 $208.29 \u2502 \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2534\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518 I am on $100/m sub. Oct 16 I switched machine Excludes work done on local and Claude Chrome plug", "context": "on $100/m Claude sub", "use_case": "tracking API usage cost", "switched_from": "", "story": "If you don't run your own LLM proxy, how do you know whats going on?"}, {"id": 802306864, "item_id": 50144179, "url": "https://news.ycombinator.com/item?id=50144179", "by": "geye1234", "date": "2026-10-18", "stance": "neutral", "quote": "I use Pi, but I think its updates and its extensions' updates are the biggest security gap I have right now.", "excerpt": "I use Pi, but I think its updates and its extensions' updates are the biggest security gap I have right now. I'm not even sure what they keep updating TBH. Any recommendations?", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "OpenCode has a security breach last that impacted 1.4% of console users"}, {"id": 802166561, "item_id": 50135410, "url": "https://news.ycombinator.com/item?id=50135410", "by": "aizk", "date": "2026-10-17", "stance": "recommend", "quote": "pi harness because it's somewhat lightweight", "excerpt": "Caveat: I get a budget of free ai tokens at work, as I work for an open source AI inference provider. My goal is to reverse engineer an entire gamecube game -- I'm pushing an existing community decomp that was stalled to the finish line, solo, with lots of AI help. This work was directly inspired by https://decomp.dev/ -- which has just shut down a few days ago because of, ironically... people like me who enjoy decom", "context": "works for an open source AI inference provider", "use_case": "agent harness", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 802095856, "item_id": 50130991, "url": "https://news.ycombinator.com/item?id=50130991", "by": "sanex", "date": "2026-10-17", "stance": "recommend", "quote": "I've a heavily customized pi that I use which let's me go hands off and not lose control", "excerpt": "I'm somewhere between 3 and 5 depending on the ticket. Fable pushed me over the edge on it but also I've a heavily customized pi that I use which let's me go hands off and not lose control. It's all about tightening those validation loops, watching where weird shit happens and trying to figure out if there was something unclear or ambiguous that let it happen. I was an engineering manager that stepped back into IC so", "context": "former engineering manager, now IC", "use_case": "AI coding agent", "switched_from": "", "story": "Ask HN: How much are developers trusting LLMs to write prod code?"}, {"id": 802082051, "item_id": 50130128, "url": "https://news.ycombinator.com/item?id=50130128", "by": "clx75", "date": "2026-10-17", "stance": "neutral", "quote": "I start up a new Pi session, switch to Deepseek 4.1 Flash", "excerpt": "Probably a week ago someone posted a thread with the title \"Why isn't the industry freaking out about DeepSeek 4.1 Flash?\" I have been intrigued about Chinese models for a long time, but never tried anything except an early Qwen. My daily driver was Luna, with occasional elevation to Terra or Sol - all on 5.6 as I found the output of 6 jarringly incomprehensible. I still had 95 USD on my OpenRouter account, so I thou", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 802068868, "item_id": 50129304, "url": "https://news.ycombinator.com/item?id=50129304", "by": "egeozcan", "date": "2026-10-17", "stance": "neutral", "quote": "All running on pi.", "excerpt": "I use Opus for planning, through API on medium thinking, it doesn't cost anything significant and the rest can be happily done with deepseek or glm flash. Using sol 6.1 for review loops at the end. All running on pi. A very complex task tends to cost like 10$ max. It's comparable to a subscription with enormous flexibility. It's a steal.", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 801868896, "item_id": 50116806, "url": "https://news.ycombinator.com/item?id=50116806", "by": "Otterly99", "date": "2026-10-16", "stance": "recommend", "quote": "a simple harness like the one from Pi is sufficient", "excerpt": "https://harnesstax.github.io/ From what I gathered, a simple harness like the one from Pi is sufficient. Similarly, I would advise against using tons of skills and keep the ones you genuienly find useful (they are just .md files so you can copy them). The only skill I use is /strategic-compact from everything claude code (ecc) because I like to read the summary before a manual compaction.", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 801848371, "item_id": 50115523, "url": "https://news.ycombinator.com/item?id=50115523", "by": "armcat", "date": "2026-10-16", "stance": "recommend", "quote": "There are lot of other amazing open source agent harnesses like OpenCode [2] and Pi [3].", "excerpt": "For Codex there is no need to decompile because unlike Claude Code, Codex is open source [1] and written in Rust. There are lot of other amazing open source agent harnesses like OpenCode [2] and Pi [3]. [1] https://github.com/openai/codex [2] https://github.com/anomalyco/opencode [3] https://github.com/earendil-works/pi", "context": "", "use_case": "agent harness", "switched_from": "", "story": "Ask HN: Has anyone decompiled Claude Code yet?"}, {"id": 801816209, "item_id": 50113513, "url": "https://news.ycombinator.com/item?id=50113513", "by": "verdverm", "date": "2026-10-16", "stance": "recommend", "quote": "next try a harness like opencode or pi with open weight models via opencode-go for $10/m", "excerpt": "yup, next try a harness like opencode or pi with open weight models via opencode-go for $10/m or make your own harness, it's easy, fun, and you learn a lot", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Ask HN: Thoughts about the Gemini Web app?"}, {"id": 801695968, "item_id": 50105998, "url": "https://news.ycombinator.com/item?id=50105998", "by": "pizzafeelsright", "date": "2026-10-15", "stance": "compare", "quote": "Pi is the most basic with the ability to add features.", "excerpt": "The harnesses all do the same thing in a general sense. They loop over chat with defined tagging/syntax that issues commands and run sub-agents that do the same. Pi is the most basic with the ability to add features. opencode also supports more features and was build as an open alternative to claude.", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 801639537, "item_id": 50102471, "url": "https://news.ycombinator.com/item?id=50102471", "by": "wg0", "date": "2026-10-15", "stance": "recommend", "quote": "That's why I first switched to Pi and then to DeepSeek harness.", "excerpt": "Performance is horrible both of the website and of the tool itself. That's why I first switched to Pi and then to DeepSeek harness.", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 801619760, "item_id": 50101235, "url": "https://news.ycombinator.com/item?id=50101235", "by": "verdverm", "date": "2026-10-15", "stance": "neutral", "quote": "I switched to Pi last week", "excerpt": "I switched to Pi last week, have not seen anything about self modification, but have seen a lot about security, the minimalism and customization by user are closest in common selling points from exiting users you cannot have both and be serious about security to be clear, I'm talking about hot reload, agent can manipulate its own environment in real time, that's asking for trouble, put some controls on that or reap t", "context": "", "use_case": "coding agent", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 801594466, "item_id": 50099654, "url": "https://news.ycombinator.com/item?id=50099654", "by": "avereveard", "date": "2026-10-15", "stance": "neutral", "quote": "manage pi command lines by pid", "excerpt": "No because it doesnt get backgrounded and I would need to hold one end of the stream at all time, and feature completedness is spotty I use the protocol claude code terminal use to talk to its server to spawn agents, opencode server protocol which is published for the same purpose, and manage pi command lines by pid", "context": "", "use_case": "agent CLI", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 801585522, "item_id": 50099095, "url": "https://news.ycombinator.com/item?id=50099095", "by": "ChaseRensberger", "date": "2026-10-15", "stance": "neutral", "quote": "i often just steal with things i like from ocv2 and pi.", "excerpt": "their client server model is a fantastic idea and the core reason i started building https://wingman.actor a year ago. unfortunately making a great harness is tricky and i often just steal with things i like from ocv2 and pi. been using ocv2 as my daily driver for a while and had very little complaints. its a great product from a great team.", "context": "", "use_case": "agent harness", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 801584176, "item_id": 50099011, "url": "https://news.ycombinator.com/item?id=50099011", "by": "usernametaken29", "date": "2026-10-15", "stance": "recommend", "quote": "I recently switched to pi which uses an order of magnitude less RAM (hundreds vs dozens)", "excerpt": "I wish they\u2019d invest more time in performance improvements over code mods. I recently switched to pi which uses an order of magnitude less RAM (hundreds vs dozens) and noticeably less CPU on my machine\u2026 OpenCode is really nice but it\u2019s about as heavy as antigravity and is substandard in features\u2026", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 801555136, "item_id": 50097196, "url": "https://news.ycombinator.com/item?id=50097196", "by": "verdverm", "date": "2026-10-14", "stance": "recommend", "quote": "I switched to Pi after experiencing OpenCode V2 about a week ago.", "excerpt": "I switched to Pi after experiencing OpenCode V2 about a week ago. Dislike the UX direction a lot. I took some effort to sift through the Pi plugin ecosystem to get my setup in a good place. I will say OpenCode has likely improved significantly from your experience. I never experienced any of that since ~May of this year when I picked up OpenCode the first time.", "context": "", "use_case": "AI coding harness", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 801554736, "item_id": 50097171, "url": "https://news.ycombinator.com/item?id=50097171", "by": "verdverm", "date": "2026-10-14", "stance": "recommend", "quote": "Pi is a good harness for vim users to consider if they want to try a TUI experience.", "excerpt": "I was a Vim user that switched to VS Code (for reasons) and now the TUIs are where my Vim personalization efforts have been redirected. Pi is a good harness for vim users to consider if they want to try a TUI experience. The Pi plugin ecosystem is a slopfest though. There are vim bindings for Pi and OC (and VS Code) because I'm never going to give those up.", "context": "former Vim user now on VS Code", "use_case": "terminal AI coding harness", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 801552592, "item_id": 50097037, "url": "https://news.ycombinator.com/item?id=50097037", "by": "verdverm", "date": "2026-10-14", "stance": "neutral", "quote": "While I just switched from OpenCode to Pi", "excerpt": "While I just switched from OpenCode to Pi, I'm still unclear on which is better in terms of handling github issues/pull-requests. I do give OC credit here, especially after the last week. Here's the Pi issue for a security feature request, derived from an opencode v2 security regression I was able to work through with the OC team. This issue links to the OpenCode ones too. https://github.com/earendil-works/pi/issues/", "context": "", "use_case": "coding agent harness", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 801546528, "item_id": 50096658, "url": "https://news.ycombinator.com/item?id=50096658", "by": "verdverm", "date": "2026-10-14", "stance": "recommend", "quote": "I've switched to Pi because I do not like the direction v2 has taken UX.", "excerpt": "I suspect heavy eyes free coding. I spent the last week helping them understand a major security regression in config loading, took two issues/pull requests. Our hunch is the Ai told them what they had was good, not enough human critical thinking. https://news.ycombinator.com/item?id=50075319 I've switched to Pi because I do not like the direction v2 has taken UX. Nor do I trust the security posturing going forward. ", "context": "", "use_case": "coding agent harness", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 801536657, "item_id": 50096041, "url": "https://news.ycombinator.com/item?id=50096041", "by": "montroser", "date": "2026-10-14", "stance": "recommend", "quote": "I switched to pi some months ago, and I have no complaints.  It just does what it says, with no shenanigans.", "excerpt": "OpenCode is a beast. Does it still try to take a disk snapshot of my working directory, even when it contains hundreds of millions of files? Does it still eat all of my memory? Does it still have a super-gigantic system prompt that wastes a solid percentage of my context window before I've even begun? I switched to pi some months ago, and I have no complaints. It just does what it says, with no shenanigans. They upda", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 801288336, "item_id": 50080521, "url": "https://news.ycombinator.com/item?id=50080521", "by": "verdverm", "date": "2026-10-13", "stance": "neutral", "quote": "I switched to pi while I go refresh my personal/custom harness", "excerpt": "I use open weights and harnesses not tied to a single vendor, I definitely do not touch anything the two US Big Ai labs puts out. I'm not interested in handing power and control of humans most significant technology to the same class of people that brought us the technologies that created our mess of a digital ecosystem. I switched to pi while I go refresh my personal/custom harness, though that may be a bit off, the", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Major security regression merged into OpenCode V2"}, {"id": 801072371, "item_id": 50067023, "url": "https://news.ycombinator.com/item?id=50067023", "by": "crossroadsguy", "date": "2026-10-13", "stance": "neutral", "quote": "Pi the least of them (not that it's worse, but because I use it for \"non coding\" purposes)", "excerpt": "So is OpenCode. More so. I use Claude Code and OpenCode almost equally and Codex slightly less and Pi the least of them (not that it's worse, but because I use it for \"non coding\" purposes). They are all bad. Even this newer breed of \"agent orchrestrating\" tools. One goes to Github and one notices, the dev is done by LLMs, issue triage is done by bots, and the commenting is done by the bots. (btw, the triage mostly m", "context": "", "use_case": "non-coding tasks", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 801011936, "item_id": 50063246, "url": "https://news.ycombinator.com/item?id=50063246", "by": "spudlyo", "date": "2026-10-12", "stance": "recommend", "quote": "I've been using Pi nearly every day for 4-5 months now", "excerpt": "I've been using Pi nearly every day for 4-5 months now, and even though it is still very much in development with new updates arriving nearly every week, it has to my mind never regressed. I've never lamented that a Pi update ever broke anything I depend on or made anything worse. I immediately adopted the new tuiMode \"fullscreen\" feature the moment it landed, and it's been exceptional. I love the way it now manages ", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 800776385, "item_id": 50048524, "url": "https://news.ycombinator.com/item?id=50048524", "by": "verdverm", "date": "2026-10-11", "stance": "recommend", "quote": "I do not like the direction they are taking and switched to pi", "excerpt": "There are a lot of breaking changes, several undocumented, and a major security regression (I was told they will merge, in discord by the dev team) https://github.com/anomalyco/opencode/pull/51337 I do not like the direction they are taking and switched to pi https://opencode.ai/v2/docs/migrate-v1/#breaking-changes If you want to stay with opencode, I would wait a while for things to settle, every plugin has to migra", "context": "", "use_case": "coding agent", "switched_from": "OpenCode", "story": "OpenCode 2.0"}, {"id": 798924704, "item_id": 49932794, "url": "https://news.ycombinator.com/item?id=49932794", "by": "nfriedly", "date": "2026-10-02", "stance": "recommend", "quote": "I really like Pi, especially for when you're running a local model and your context is particularly limited.", "excerpt": "I really like Pi, especially for when you're running a local model and your context is particularly limited. But even when running a big model, I think building up your environment to be exactly what you want can be really powerful. When I was first getting started and didn't realize that ollama defaults to incredibly tiny amounts of context, Pi was the only harness that actually worked because the rest of them used ", "context": "", "use_case": "coding harness for local models", "switched_from": "", "story": "Pi 1.0"}, {"id": 798923888, "item_id": 49932743, "url": "https://news.ycombinator.com/item?id=49932743", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "I've daily driven Pi for 9+ months and I recommend pi a lot to friends.", "excerpt": "Posted this elsewhere in this thread. I've daily driven Pi for 9+ months and I recommend pi a lot to friends. I think switching to pi from claude code is pretty daunting. The pi extension system is excellent. But which extensions should you use? And honestly a lot of the most popular extensions to me feel bloated / weird / vibe-coded. Here's my minimal set up: https://github.com/richardgill/pi-extensions/blob/main/PI", "context": "daily driver for 9+ months", "use_case": "coding agent", "switched_from": "Claude Code", "story": "Pi 1.0"}, {"id": 798922816, "item_id": 49932676, "url": "https://news.ycombinator.com/item?id=49932676", "by": "richardgill88", "date": "2026-10-02", "stance": "recommend", "quote": "Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-coding projects than seriously maintained tools.", "excerpt": "Long-time pi user here. If you're thinking of trying out pi but feel daunted by finding the right plugins to get started, I wrote about my minimal pi setup here: https://github.com/richardgill/pi-extensions/blob/main/PI_SETUP.md It's still pretty minimal and brings back just enough stuff that I don't miss Claude Code. Pi's plugins are its strength, but also its weakness, because most feel more like personal vibe-codi", "context": "Long-time pi user", "use_case": "AI coding agent", "switched_from": "", "story": "Pi 1.0"}, {"id": 798898528, "item_id": 49931158, "url": "https://news.ycombinator.com/item?id=49931158", "by": "eloisant", "date": "2026-10-02", "stance": "recommend", "quote": "I switched from OpenCode to Pi and never looked back.", "excerpt": "Yes, I switched from OpenCode to Pi and never looked back. What's great with Pi is that it's very easy to extend, because it knows its own doc. So you tell it \"implement a plugin that does this\" and it does it directly in its own folder. That means I have a plan mode that works exactly the way I want, I have a hook that cleans up added comments after each change, I can ask \"pull this github PR and assess the comments", "context": "", "use_case": "AI coding agent", "switched_from": "OpenCode", "story": "Pi 1.0"}, {"id": 798881744, "item_id": 49930109, "url": "https://news.ycombinator.com/item?id=49930109", "by": "by2kaapi", "date": "2026-10-02", "stance": "recommend", "quote": "But loved the minimalism, extensibility! It's a very capable harness.", "excerpt": "Switched to pi to avoid having to keep skills/MCP configs in sync across agents like claude and codex to have consistent experience. But loved the minimalism, extensibility! It's a very capable harness. Have my own extensions for MCP, subagents, ui, history management etc. overall i feel my workflow has got a lot better. Love pi!", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798868401, "item_id": 49929275, "url": "https://news.ycombinator.com/item?id=49929275", "by": "pettijohn", "date": "2026-10-02", "stance": "recommend", "quote": "It's been so easy to bend into what I need.", "excerpt": "I use PI WEB (https://pi-web.dev, there's more than one with that name) to orchestrate remotely. Pi lets me run local models (currently Qwen 3.8 Flash Next on Strix Halo 128GB) and flagship models (with subscription auth, not API pricing) side by side. I typically ask GPT 6.1 Sol to review requirements and then spawn a subsession with Qwen, review Qwen's work, ask Qwen to fix. About 95% success with a single pass lik", "context": "runs local models on Strix Halo 128GB", "use_case": "mixing local and flagship models", "switched_from": "", "story": "Pi 1.0"}, {"id": 798849136, "item_id": 49928071, "url": "https://news.ycombinator.com/item?id=49928071", "by": "lo5", "date": "2026-10-01", "stance": "recommend", "quote": "It feels light and nimble.", "excerpt": "I use Pi with only a couple of extensions and using less and less of CC and Codex everyday. It feels light and nimble. Codemode from yesterday's release made it even better. I wish I could use my Claude subscription via Pi.", "context": "", "use_case": "AI coding agent", "switched_from": "Claude Code", "story": "Pi 1.0"}, {"id": 798842672, "item_id": 49927667, "url": "https://news.ycombinator.com/item?id=49927667", "by": "darklinear", "date": "2026-10-01", "stance": "recommend", "quote": "I find that it does better than the native codex harness", "excerpt": "I use pi with a ChatGPT subscription. I find that it does better than the native codex harness: - it's very token efficient. I've had whole profiling + refactoring sessions finish in <20k tokens, whereas in codex you can get close to that with just all the cruft OpenAI loads by default before you even send your own messages - the token efficiency means turns are a lot snappier - fine-grained context control by defaul", "context": "", "use_case": "profiling and refactoring with ChatGPT subscription", "switched_from": "Codex", "story": "Pi 1.0"}]}