{"slug": "oh-my-pi", "name": "oh-my-pi", "category": "ai-coding", "category_name": "AI coding tools", "description": "A batteries-included terminal AI coding agent harness built on the minimalist Pi agent.", "recommenders": 36, "commenters": 56, "warners": 2, "score": 35.928, "rising": 18.76, "first": "2026-02-24", "latest": "2026-10-27", "why": "HN commenters recommend it as a batteries-included Pi distribution that works well out of the box, with flexible per-role model assignment, a built-in plan mode, subagents and LSP-based edits.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 45, "neutral": 19, "compare": 3, "warn": 2}, "spark": [0, 0, 0, 1, 0, 1, 0, 6, 9, 13, 7, 8], "listed": true, "loved": ["Batteries included", "Multi-model roles"], "caveat": "Context overhead", "short_quote": "I can strongly recommend omp.sh if you just want to be up and running without hassle and a far better (for me) developer experience.", "icon": true, "domain": null, "github_repo": "github.com/can1357/oh-my-pi", "official_url": "https://github.com/can1357/oh-my-pi", "pricing": "open source", "aliases": ["oh-my-pi", "OMP", "omp", "Oh My Pi", "ohmypi", "omp.sh", "Oh-My-Pi", "OMP (oh-my-pi)", "omp (oh my pi)"], "verdict": [{"text": "HN commenters recommend it as a batteries-included Pi distribution that works well out of the box,", "cites": [{"id": 798853984, "item_id": 49928374, "url": "https://news.ycombinator.com/item?id=49928374", "by": "CharlesW", "date": "2026-10-01", "stance": "recommend", "quote": "I recommend the batteries-included Pi distribution oh-my-pi (omp.sh) over Pi and OpenCode.", "excerpt": "As someone similarly invested in Claude Code and initially reluctant to try other harnesses, I recommend the batteries-included Pi distribution oh-my-pi (omp.sh) over Pi and OpenCode. I switch between Claude Code, omp, and Codex regularly, and it feels fine.", "context": "heavily invested in Claude Code", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798853040, "item_id": 49928315, "url": "https://news.ycombinator.com/item?id=49928315", "by": "CharlesW", "date": "2026-10-01", "stance": "recommend", "quote": "I can strongly recommend omp.sh if you just want to be up and running without hassle and a far better (for me) developer experience.", "excerpt": "That the \"batteries included\" bit, but it's not \"burning tokens\" in the sense that using Pi with the same batteries would (of course) use just as many. I can strongly recommend omp.sh if you just want to be up and running without hassle and a far better (for me) developer experience.", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 800071152, "item_id": 50004447, "url": "https://news.ycombinator.com/item?id=50004447", "by": "jeremyjh", "date": "2026-10-08", "stance": "recommend", "quote": "I use oh-my-pi (omp.sh) - mostly with stock settings and skills but its a \"fully loaded\" harness that you don't really have to add anything to.", "excerpt": "I use oh-my-pi (omp.sh) - mostly with stock settings and skills but its a \"fully loaded\" harness that you don't really have to add anything to. I do change a couple of things: I set it to prefer subagents, and I enable rewind. It is crazy that rewind is not a default, it saves a TON of context - if the agent goes down a crazy path that burns a lot of tokens it can rewind to an earlier checkpoint with the exploration ", "context": "", "use_case": "AI coding agent harness", "switched_from": "", "story": "Claude Haiku 5.5"}, {"id": 781580480, "item_id": 48848780, "url": "https://news.ycombinator.com/item?id=48848780", "by": "roger_", "date": "2026-07-09", "stance": "recommend", "quote": "Currently using Oh My Pi (https://github.com/can1357/oh-my-pi) and appreciate the batteries included approach.", "excerpt": "Currently using Oh My Pi (https://github.com/can1357/oh-my-pi) and appreciate the batteries included approach. From my limited time using pure Pi, I found quite a few of the plugins lacking and had no desire to upgrade/fix and maintain them myself. I know others feel differently though. I like the idea of keep Pi minimal but having \u201cofficial\u201d, high quality optional plugins to make it more usable.", "context": "", "use_case": "coding agent config", "switched_from": "Pi", "story": "Opinionated and Easy Pi.dev Configuration"}, {"id": 787220130, "item_id": 49201258, "url": "https://news.ycombinator.com/item?id=49201258", "by": "eli", "date": "2026-08-06", "stance": "recommend", "quote": "OpenCode or oh-my-pi might make more sense if you just want a batteries-included agent.", "excerpt": "I use https://pi.dev/ which works fine out of the box but is fairly minimal and intended to be customized. There are many extensions. OpenCode or oh-my-pi might make more sense if you just want a batteries-included agent. You can also make Claude Code work with other models without too much work, but I think that's asking for headaches.", "context": "", "use_case": "CLI coding agent", "switched_from": "", "story": "Qwen3.8 Max now ranked as the best overall model by agentic index"}]}, {"text": " with flexible per-role model assignment, a built-in plan mode, subagents and LSP-based edits.", "cites": [{"id": 784478160, "item_id": 49029885, "url": "https://news.ycombinator.com/item?id=49029885", "by": "greggh", "date": "2026-07-23", "stance": "recommend", "quote": "Stock OMP. When you add a provider and then /model you can choose a model and choose what roles it will take.", "excerpt": "Stock OMP. When you add a provider and then /model you can choose a model and choose what roles it will take. Do that for each role it has. Then I use /plan when I need to be planning and not writing to any files. Take the advice the UI gives you where it tells you to add the word orchestrate into your prompts when you want to make sure it uses its todo/tasks and sub-agents.", "context": "", "use_case": "agentic coding with planning", "switched_from": "", "story": "Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA"}, {"id": 796936433, "item_id": 49808527, "url": "https://news.ycombinator.com/item?id=49808527", "by": "jeremyjh", "date": "2026-09-22", "stance": "recommend", "quote": "In OMP I have Sol high as the orchestrator, Sol max as Planner & Reviewer, Luna max as task/coder. Very good setup.", "excerpt": "You can get a lot of Codex usage out of that same sub on top of ChatGPT usage. Its a really good value and you can use that sub in any harness. In OMP I have Sol high as the orchestrator, Sol max as Planner & Reviewer, Luna max as task/coder. Very good setup. I'm on pro now and there are weekends when I use half a week's usage but I'll have 5 or 6 sessions going at once for many hours each day.", "context": "on Pro plan", "use_case": "multi-agent coding harness", "switched_from": "", "story": "GPT-6 Sol and Luna"}, {"id": 797642432, "item_id": 49852652, "url": "https://news.ycombinator.com/item?id=49852652", "by": "spamizbad", "date": "2026-09-26", "stance": "recommend", "quote": "I greatly enjoy using the plan mode that comes baked into omp (oh my pi)", "excerpt": "It's been a while (well, 8 months) since I used Claude Code, but what I can say is that I greatly enjoy using the plan mode that comes baked into omp (oh my pi), especially for tasks around firmware development. I recognize that has a very different workflow than webapps though.", "context": "firmware developer", "use_case": "firmware development planning", "switched_from": "Claude Code", "story": "Plan mode is dead"}, {"id": 777826080, "item_id": 48614130, "url": "https://news.ycombinator.com/item?id=48614130", "by": "Bnjoroge", "date": "2026-06-20", "stance": "recommend", "quote": "I use OMP. Was on Pi but switched over because I like OMP's subagents implementation.", "excerpt": "I use OMP. Was on Pi but switched over because I like OMP's subagents implementation. Use 5.5/4.8 for coming up with a very detailed plan, ds4pro/mimo for implementation of easy features, 5.5/4.6 for larger/harder features, and 5.5/4.6 for a review. I use them in onorca.dev and sometimes use Zed as well.", "context": "", "use_case": "AI coding agent", "switched_from": "Pi", "story": "Ask HN: Do you use Claude Code, Codex, or something else?"}, {"id": 787732000, "item_id": 49233250, "url": "https://news.ycombinator.com/item?id=49233250", "by": "latchkey", "date": "2026-08-09", "stance": "recommend", "quote": "i just switched to it from codex... and i feel like i just leveled up. what sells it for me is that it uses LSP's to edit the files.", "excerpt": "i was avoiding the harness wars, but a friend recommended it and i just switched to it from codex... and i feel like i just leveled up. what sells it for me is that it uses LSP's to edit the files. so much faster and more accurate edits. use it with ACP through Zed and you get rid of all the TUI graphics nonsense...", "context": "", "use_case": "coding agent harness", "switched_from": "Codex", "story": "I Wanted to Own the Harness. Then Codex Desktop Won"}]}, {"text": " The most common caveats are the context its extra structure uses up", "cites": [{"id": 798837841, "item_id": 49927365, "url": "https://news.ycombinator.com/item?id=49927365", "by": "razster", "date": "2026-10-01", "stance": "warn", "quote": "I've tried Oh-my-pi but it's too heavy for me and eats up 6% of my context on start.", "excerpt": "I use my Pi to organize my graphic design assets as well as update documentation for client projects. It also is my daily driver for a story I'm working on and thanks to a ton of great skills created it has become more powerful. Local model is Qwen3.6 35B-A3B and Qwen3.8 27B UD. I love the compressing function, keeps on pushing. I've tried Oh-my-pi but it's too heavy for me and eats up 6% of my context on start. Also", "context": "", "use_case": "", "switched_from": "", "story": "Pi 1.0"}, {"id": 798869568, "item_id": 49929348, "url": "https://news.ycombinator.com/item?id=49929348", "by": "rpdillon", "date": "2026-10-02", "stance": "recommend", "quote": "OMP provides a significant amount of structure to the work.", "excerpt": "OMP provides a significant amount of structure to the work. That costs some tokens in terms of the system prompt. When I use these systems, I balance the return of the extra guidance in the system prompt with the cost of the context. You've talked about how much context it takes, but you haven't talked about any difference in behavior. omp offers some nice tools, like /shake, to manage context efficiently and offset ", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}]}, {"text": " and a steep learning curve that suits people who like to tinker.", "cites": [{"id": 783830192, "item_id": 48989387, "url": "https://news.ycombinator.com/item?id=48989387", "by": "217", "date": "2026-07-21", "stance": "recommend", "quote": "most advanced, configurable and capable harness of all time", "excerpt": "https://github.com/can1357/oh-my-pi https://stencil.so/blog/prewalk most advanced, configurable and capable harness of all time, and i use both 200$ subs with it. Has a huge learning curve though so if you're not into that i'd go with whatever you already use the most", "context": "pays for two $200 subscriptions", "use_case": "coding agent harness", "switched_from": "", "story": "Ask HN: Claude Code or Codex?"}, {"id": 788984993, "item_id": 49311562, "url": "https://news.ycombinator.com/item?id=49311562", "by": "217", "date": "2026-08-15", "stance": "recommend", "quote": "omp.sh if youre into figuring out and tinkering with the best tools available", "excerpt": "Just make skills for things you're doing repeatedly / often correcting the model on. If you are just starting i highly recommend trying out codex desktop app (it's good and just works, especially if you are on the plan that has 5.6 sol) or omp.sh if youre into figuring out and tinkering with the best tools available", "context": "", "use_case": "tinkering with agent tooling", "switched_from": "", "story": "Ask HN: What agent skills are necessary?"}]}], "praise": [{"label": "Batteries included", "summary": "Commenters say it adds useful tools to Pi's minimal base, so it is usable without assembling extensions yourself.", "quotes": [{"id": 798853984, "item_id": 49928374, "url": "https://news.ycombinator.com/item?id=49928374", "by": "CharlesW", "date": "2026-10-01", "stance": "recommend", "quote": "I recommend the batteries-included Pi distribution oh-my-pi (omp.sh) over Pi and OpenCode.", "excerpt": "As someone similarly invested in Claude Code and initially reluctant to try other harnesses, I recommend the batteries-included Pi distribution oh-my-pi (omp.sh) over Pi and OpenCode. I switch between Claude Code, omp, and Codex regularly, and it feels fine.", "context": "heavily invested in Claude Code", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 781580480, "item_id": 48848780, "url": "https://news.ycombinator.com/item?id=48848780", "by": "roger_", "date": "2026-07-09", "stance": "recommend", "quote": "Currently using Oh My Pi (https://github.com/can1357/oh-my-pi) and appreciate the batteries included approach.", "excerpt": "Currently using Oh My Pi (https://github.com/can1357/oh-my-pi) and appreciate the batteries included approach. From my limited time using pure Pi, I found quite a few of the plugins lacking and had no desire to upgrade/fix and maintain them myself. I know others feel differently though. I like the idea of keep Pi minimal but having \u201cofficial\u201d, high quality optional plugins to make it more usable.", "context": "", "use_case": "coding agent config", "switched_from": "Pi", "story": "Opinionated and Easy Pi.dev Configuration"}, {"id": 798868880, "item_id": 49929305, "url": "https://news.ycombinator.com/item?id=49929305", "by": "rpdillon", "date": "2026-10-02", "stance": "recommend", "quote": "I ended up choosing omp as my daily driver", "excerpt": "Yeah, saying it kills the minimalist bit misses the point of omp being a separate project. It took a minimalist-base of pi and added tools that make it usable out of the box. Hence batteries. One example here is that it added MCP extensions to make the harness usable with MCP servers long before Pi figured out that that was the right move. A lot can be said about development of MCP over that timeframe, but I apprecia", "context": "Long-time Linux and vanilla Emacs user", "use_case": "AI coding harness", "switched_from": "Pi", "story": "Pi 1.0"}]}, {"label": "Multi-model roles", "summary": "Users like adding providers and assigning different models to roles such as orchestrator, planner, advisor or coder.", "quotes": [{"id": 784478160, "item_id": 49029885, "url": "https://news.ycombinator.com/item?id=49029885", "by": "greggh", "date": "2026-07-23", "stance": "recommend", "quote": "Stock OMP. When you add a provider and then /model you can choose a model and choose what roles it will take.", "excerpt": "Stock OMP. When you add a provider and then /model you can choose a model and choose what roles it will take. Do that for each role it has. Then I use /plan when I need to be planning and not writing to any files. Take the advice the UI gives you where it tells you to add the word orchestrate into your prompts when you want to make sure it uses its todo/tasks and sub-agents.", "context": "", "use_case": "agentic coding with planning", "switched_from": "", "story": "Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA"}, {"id": 796936433, "item_id": 49808527, "url": "https://news.ycombinator.com/item?id=49808527", "by": "jeremyjh", "date": "2026-09-22", "stance": "recommend", "quote": "In OMP I have Sol high as the orchestrator, Sol max as Planner & Reviewer, Luna max as task/coder. Very good setup.", "excerpt": "You can get a lot of Codex usage out of that same sub on top of ChatGPT usage. Its a really good value and you can use that sub in any harness. In OMP I have Sol high as the orchestrator, Sol max as Planner & Reviewer, Luna max as task/coder. Very good setup. I'm on pro now and there are weekends when I use half a week's usage but I'll have 5 or 6 sessions going at once for many hours each day.", "context": "on Pro plan", "use_case": "multi-agent coding harness", "switched_from": "", "story": "GPT-6 Sol and Luna"}, {"id": 788746657, "item_id": 49296666, "url": "https://news.ycombinator.com/item?id=49296666", "by": "ljosifov", "date": "2026-08-14", "stance": "recommend", "quote": "Lately I use Zai in omp (oh-my-pi).", "excerpt": "I've used GLM-s the longest with Claude Code and their Anthropic supplied endpoint. As per their docs $ ANTHROPIC_BASE_URL=\"https://api.z.ai/api/anthropic\" ANTHROPIC_AUTH_TOKEN=\"zai-api-key\" claude --dangerously-skip-permissions Lately I use Zai in omp (oh-my-pi). It's listed built-in provider can be selected without configs shenanigans. Fits in the overall setup e.g. can select GLM-5.2 (now 5.3), and assign it role ", "context": "", "use_case": "GLM harness with roles", "switched_from": "Claude Code", "story": "GLM-5.3: Frontier Coding with Emergent Cyber Capabilities"}]}, {"label": "Configurable, extensible", "summary": "Users describe it as highly configurable and extensible, including custom guardrails and installed skills.", "quotes": [{"id": 798869568, "item_id": 49929348, "url": "https://news.ycombinator.com/item?id=49929348", "by": "rpdillon", "date": "2026-10-02", "stance": "recommend", "quote": "OMP provides a significant amount of structure to the work.", "excerpt": "OMP provides a significant amount of structure to the work. That costs some tokens in terms of the system prompt. When I use these systems, I balance the return of the extra guidance in the system prompt with the cost of the context. You've talked about how much context it takes, but you haven't talked about any difference in behavior. omp offers some nice tools, like /shake, to manage context efficiently and offset ", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 762216208, "item_id": 47638513, "url": "https://news.ycombinator.com/item?id=47638513", "by": "syl5x", "date": "2026-04-04", "stance": "recommend", "quote": "The reason that I use ohmypi is because I can extend it and put guardrails specific for our company and myself", "excerpt": "I am actively using ohmypi harness which is based on pi-mono which I believe is within OpenClaw, I don't personally use OpenClaw but I suspect that I will be affected. The reason that I use ohmypi is because I can extend it and put guardrails specific for our company and myself (those are different from SKILLs and more sophisticated than the hooks) + I like the ability to start \"tasks\" with faster models like gpt5.4-", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Anthropic no longer allowing Claude Code subscriptions to be used with OpenClaw"}, {"id": 787782080, "item_id": 49236380, "url": "https://news.ycombinator.com/item?id=49236380", "by": "g42gregory", "date": "2026-08-09", "stance": "recommend", "quote": "I use OMP with Matt Pocock skills installed and it works great.", "excerpt": "I use OMP with Matt Pocock skills installed and it works great. Code quality is better and it uses less tokens. I didn't extend it much, but it certainly is very extendable.", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "Having fun with oh my pi, DeepSeek-V4-Flash, GPT-5.6 Luna and Antigravity CLI"}]}, {"label": "Plan mode, subagents", "summary": "Users like its built-in plan mode and its subagents implementation.", "quotes": [{"id": 797642432, "item_id": 49852652, "url": "https://news.ycombinator.com/item?id=49852652", "by": "spamizbad", "date": "2026-09-26", "stance": "recommend", "quote": "I greatly enjoy using the plan mode that comes baked into omp (oh my pi)", "excerpt": "It's been a while (well, 8 months) since I used Claude Code, but what I can say is that I greatly enjoy using the plan mode that comes baked into omp (oh my pi), especially for tasks around firmware development. I recognize that has a very different workflow than webapps though.", "context": "firmware developer", "use_case": "firmware development planning", "switched_from": "Claude Code", "story": "Plan mode is dead"}, {"id": 802256306, "item_id": 50141019, "url": "https://news.ycombinator.com/item?id=50141019", "by": "pimeys", "date": "2026-10-18", "stance": "recommend", "quote": "I use the omp plan mode and iterate on the text until I'm happy with it.", "excerpt": "I also used to plan with more expensive models like you did, but now I just do planning with DeepSeek too and don't see much difference. I use the omp plan mode and iterate on the text until I'm happy with it. Then clear context, give the plan to DeepSeek again for implementation. If it's mission-critical code, do a security review with GLM 5.3. Otherwise go manually through the code and ask for changes if needed. Wo", "context": "", "use_case": "plan mode for coding", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 777826080, "item_id": 48614130, "url": "https://news.ycombinator.com/item?id=48614130", "by": "Bnjoroge", "date": "2026-06-20", "stance": "recommend", "quote": "I use OMP. Was on Pi but switched over because I like OMP's subagents implementation.", "excerpt": "I use OMP. Was on Pi but switched over because I like OMP's subagents implementation. Use 5.5/4.8 for coming up with a very detailed plan, ds4pro/mimo for implementation of easy features, 5.5/4.6 for larger/harder features, and 5.5/4.6 for a review. I use them in onorca.dev and sometimes use Zed as well.", "context": "", "use_case": "AI coding agent", "switched_from": "Pi", "story": "Ask HN: Do you use Claude Code, Codex, or something else?"}]}, {"label": "Beats other harnesses", "summary": "Several commenters say it performs or behaves better than Claude Code, OpenCode and other harnesses they have tried.", "quotes": [{"id": 792722176, "item_id": 49545136, "url": "https://news.ycombinator.com/item?id=49545136", "by": "benw214", "date": "2026-09-03", "stance": "recommend", "quote": "OMP i would say comes out on this prompt the best out of all the harnesses i've tried.", "excerpt": "I'm not sure i agree with these. I've been testing a simple prompt with different model and harness combinations to work out which one produces best results. I do this in '/goal' mode. Prompt - \"Build a single-page Three.js sci-fi hangar with hovering drones, animated warning lights, emissive runway strips, and subtle volumetric-style fog planes. Include drone formation toggle and cinematic camera path. Output one se", "context": "testing harness and model combinations", "use_case": "coding agent harness", "switched_from": "", "story": "Show HN: FrontierHarness Eval \u2013 9 harness, same model, cost per pass varies 17x"}, {"id": 789973904, "item_id": 49373369, "url": "https://news.ycombinator.com/item?id=49373369", "by": "rurban", "date": "2026-08-20", "stance": "recommend", "quote": "Behaves better than opencode, and I love their /collab remote control.", "excerpt": "This sound exactly like the new DeepSeek Harness cli. But I switched to oh-my-pi and crof.ai recently (cheap deepseek). Behaves better than opencode, and I love their /collab remote control.", "context": "", "use_case": "coding agent harness", "switched_from": "opencode", "story": "fx :Tiny, open, native coding agent."}, {"id": 804099107, "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"}]}, {"label": "LSP-based editing", "summary": "Users say editing files through LSPs gives faster and more accurate edits.", "quotes": [{"id": 787732000, "item_id": 49233250, "url": "https://news.ycombinator.com/item?id=49233250", "by": "latchkey", "date": "2026-08-09", "stance": "recommend", "quote": "i just switched to it from codex... and i feel like i just leveled up. what sells it for me is that it uses LSP's to edit the files.", "excerpt": "i was avoiding the harness wars, but a friend recommended it and i just switched to it from codex... and i feel like i just leveled up. what sells it for me is that it uses LSP's to edit the files. so much faster and more accurate edits. use it with ACP through Zed and you get rid of all the TUI graphics nonsense...", "context": "", "use_case": "coding agent harness", "switched_from": "Codex", "story": "I Wanted to Own the Harness. Then Codex Desktop Won"}, {"id": 793257537, "item_id": 49578596, "url": "https://news.ycombinator.com/item?id=49578596", "by": "latchkey", "date": "2026-09-05", "stance": "recommend", "quote": "Using omp helps a lot too, the fact that it uses an LSP under the covers is a game changer.", "excerpt": "I'm addicted. This is the most fun I've had coding in decades. I feel like I've got superpowers and the limit is just how quickly I can prompt out my ideas. Those 500m are for coding. Started with $20 and now I'm on the $200/mo codex plan. Resets are crazy... free tokens, why not take advantage of all that VC money? Using omp helps a lot too, the fact that it uses an LSP under the covers is a game changer. I'm buildi", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Ask HN: Who is using MCP in production?"}]}], "caveats": [{"label": "Context overhead", "summary": "Its extra structure costs system-prompt tokens; one user saw it use 6% of context at startup.", "quotes": [{"id": 798837841, "item_id": 49927365, "url": "https://news.ycombinator.com/item?id=49927365", "by": "razster", "date": "2026-10-01", "stance": "warn", "quote": "I've tried Oh-my-pi but it's too heavy for me and eats up 6% of my context on start.", "excerpt": "I use my Pi to organize my graphic design assets as well as update documentation for client projects. It also is my daily driver for a story I'm working on and thanks to a ton of great skills created it has become more powerful. Local model is Qwen3.6 35B-A3B and Qwen3.8 27B UD. I love the compressing function, keeps on pushing. I've tried Oh-my-pi but it's too heavy for me and eats up 6% of my context on start. Also", "context": "", "use_case": "", "switched_from": "", "story": "Pi 1.0"}, {"id": 798869568, "item_id": 49929348, "url": "https://news.ycombinator.com/item?id=49929348", "by": "rpdillon", "date": "2026-10-02", "stance": "recommend", "quote": "OMP provides a significant amount of structure to the work.", "excerpt": "OMP provides a significant amount of structure to the work. That costs some tokens in terms of the system prompt. When I use these systems, I balance the return of the extra guidance in the system prompt with the cost of the context. You've talked about how much context it takes, but you haven't talked about any difference in behavior. omp offers some nice tools, like /shake, to manage context efficiently and offset ", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}]}, {"label": "Learning curve", "summary": "It has a huge learning curve and suits people who enjoy tinkering.", "quotes": [{"id": 783830192, "item_id": 48989387, "url": "https://news.ycombinator.com/item?id=48989387", "by": "217", "date": "2026-07-21", "stance": "recommend", "quote": "most advanced, configurable and capable harness of all time", "excerpt": "https://github.com/can1357/oh-my-pi https://stencil.so/blog/prewalk most advanced, configurable and capable harness of all time, and i use both 200$ subs with it. Has a huge learning curve though so if you're not into that i'd go with whatever you already use the most", "context": "pays for two $200 subscriptions", "use_case": "coding agent harness", "switched_from": "", "story": "Ask HN: Claude Code or Codex?"}, {"id": 788984993, "item_id": 49311562, "url": "https://news.ycombinator.com/item?id=49311562", "by": "217", "date": "2026-08-15", "stance": "recommend", "quote": "omp.sh if youre into figuring out and tinkering with the best tools available", "excerpt": "Just make skills for things you're doing repeatedly / often correcting the model on. If you are just starting i highly recommend trying out codex desktop app (it's good and just works, especially if you are on the plan that has 5.6 sol) or omp.sh if youre into figuring out and tinkering with the best tools available", "context": "", "use_case": "tinkering with agent tooling", "switched_from": "", "story": "Ask HN: What agent skills are necessary?"}]}, {"label": "ACP integration issues", "summary": "One user ran into many issues using it over ACP in Zed and went back to the terminal.", "quotes": [{"id": 787031474, "item_id": 49189467, "url": "https://news.ycombinator.com/item?id=49189467", "by": "tamimio", "date": "2026-08-05", "stance": "warn", "quote": "when I tried ACP with opencode and even omp, it gave me many issues that it was easier just to use the terminal directly", "excerpt": "I have been using zed for almost 2ish years now, great overall, but last month when I tried ACP with opencode and even omp, it gave me many issues that it was easier just to use the terminal directly, oh, and I couldn\u2019t select the model that time not sure if it\u2019s fixed now.", "context": "", "use_case": "agent in Zed via ACP", "switched_from": "", "story": "Zed DeltaDB"}]}, {"label": "Frequent updates", "summary": "It updates very often, so one user runs it sandboxed as a safeguard.", "quotes": [{"id": 778668480, "item_id": 48666780, "url": "https://news.ycombinator.com/item?id=48666780", "by": "gandreani", "date": "2026-06-24", "stance": "recommend", "quote": "My harness is https://github.com/can1357/oh-my-pi. I trust it...enough.", "excerpt": "I use both the openai subscription and the opencode go subscription. I use the go subscription for my personal work and the openai subscription for my consulting work. The differences between the models are minimal, but I usually stick with gpt-5.4-mini, gpt-5.4, mimo-pro-2.5, deepseek-v4-pro. These latter ones have way more usage than even using 5.4-mini so I tend to use them in personal projects for that reason. My", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "GLM-5.2 is the step change for open agents"}]}, {"label": "No tool limiting", "summary": "One user wished they could restrict tool access when using midsized local models.", "quotes": [{"id": 781722432, "item_id": 48857652, "url": "https://news.ycombinator.com/item?id=48857652", "by": "rao-v", "date": "2026-07-10", "stance": "recommend", "quote": "I like that Oh-my-pi actually adds new value over extensions", "excerpt": "I'm exploring the pi based options now and I like that Oh-my-pi actually adds new value over extensions (the advisor and stream interruption hooks are pretty clever). I just wish it had a way for me to downlimit tool access (I love midsized local models, and I'd love to enable only 30% of the power for some usecases).", "context": "uses midsized local models", "use_case": "coding agent harness", "switched_from": "", "story": "Opinionated and Easy Pi.dev Configuration"}]}, {"label": "Claude ToS risk", "summary": "Its native Claude subscription support works by impersonating Claude Code, which a commenter notes is a ToS violation.", "quotes": [{"id": 796919072, "item_id": 49807442, "url": "https://news.ycombinator.com/item?id=49807442", "by": "glub", "date": "2026-09-22", "stance": "recommend", "quote": "I have been using oh-my-pi with 3 claude subs for the past few months without any issues.", "excerpt": "Not natively, as it's still a ToS violation and adding that in pi would go against pi principles, but there are many plugins/proxies that make it work. oh-my-pi supports it natively (again, still a ToS violation), by impersonating claude code's fingerprints. I have been using oh-my-pi with 3 claude subs for the past few months without any issues. Even native server-side OAI/ANT compaction works out of the box.", "context": "", "use_case": "using Claude subscriptions in a harness", "switched_from": "", "story": "GPT-6 Sol and Luna"}]}], "who": [{"by": "LaurensBER", "context": "Windows user on a corporate device", "url": "https://news.ycombinator.com/item?id=50095490", "date": "2026-10-14"}, {"by": "broodbucket", "context": "15+ year Arch user", "url": "https://news.ycombinator.com/item?id=49932745", "date": "2026-10-02"}, {"by": "rpdillon", "context": "Long-time Linux and vanilla Emacs user", "url": "https://news.ycombinator.com/item?id=49929305", "date": "2026-10-02"}, {"by": "CharlesW", "context": "heavily invested in Claude Code", "url": "https://news.ycombinator.com/item?id=49928374", "date": "2026-10-01"}, {"by": "spamizbad", "context": "firmware developer", "url": "https://news.ycombinator.com/item?id=49852652", "date": "2026-09-26"}, {"by": "jeremyjh", "context": "on Pro plan", "url": "https://news.ycombinator.com/item?id=49808527", "date": "2026-09-22"}, {"by": "benw214", "context": "testing harness and model combinations", "url": "https://news.ycombinator.com/item?id=49545136", "date": "2026-09-03"}, {"by": "pimeys", "context": "heavy work user of SOTA models", "url": "https://news.ycombinator.com/item?id=49407637", "date": "2026-08-23"}, {"by": "217", "context": "pays for two $200 subscriptions", "url": "https://news.ycombinator.com/item?id=48989387", "date": "2026-07-21"}, {"by": "rao-v", "context": "uses midsized local models", "url": "https://news.ycombinator.com/item?id=48857652", "date": "2026-07-10"}, {"by": "trollbridge", "context": "small dev team of 6", "url": "https://news.ycombinator.com/item?id=48648590", "date": "2026-06-23"}, {"by": "rurban", "context": "mostly C, C++, perl and python developer", "url": "https://news.ycombinator.com/item?id=48423534", "date": "2026-06-06"}], "tags": ["for Pi users wanting more", "for multi-model setups", "for DeepSeek and open models", "for tinkerers", "works with Zed"], "use_cases": ["coding agent harness", "ai coding harness", "agent harness", "ai coding agent", "coding harness", "ai coding"], "compared": [{"slug": "pi", "name": "Pi", "co": 30, "switched_from": 5, "switched_to": 3}, {"slug": "opencode", "name": "OpenCode", "co": 20, "switched_from": 5, "switched_to": 0}, {"slug": "claude-code", "name": "Claude Code", "co": 20, "switched_from": 4, "switched_to": 0}, {"slug": "codex", "name": "Codex", "co": 21, "switched_from": 1, "switched_to": 0}, {"slug": "claude-opus", "name": "Claude Opus", "co": 12, "switched_from": 0, "switched_to": 0}, {"slug": "deepseek", "name": "DeepSeek", "co": 9, "switched_from": 0, "switched_to": 0}, {"slug": "gpt-sol", "name": "GPT Sol", "co": 8, "switched_from": 0, "switched_to": 0}, {"slug": "deepseek-v4-flash", "name": "DeepSeek V4 Flash", "co": 7, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=49926069", "title": "Pi 1.0", "mentions": 10}, {"url": "https://news.ycombinator.com/item?id=48847407", "title": "Opinionated and Easy Pi.dev Configuration", "mentions": 3}, {"url": "https://news.ycombinator.com/item?id=49896586", "title": "GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=49805509", "title": "GPT-6 Sol and Luna", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=49760187", "title": "Claude Code now reads AGENTS.md if there is no Claude.md", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=49367350", "title": "Anthropic Refuses to Support Agents.md", "mentions": 2}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {"recommend": 1}, "2026-03": {}, "2026-04": {"recommend": 1}, "2026-05": {}, "2026-06": {"neutral": 5, "recommend": 6}, "2026-07": {"recommend": 9, "neutral": 4}, "2026-08": {"recommend": 13, "warn": 1, "neutral": 1, "compare": 1}, "2026-09": {"recommend": 7, "neutral": 4}, "2026-10": {"warn": 1, "recommend": 8, "compare": 2, "neutral": 5}}, "quotes": [{"id": 804099107, "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": 803567505, "item_id": 50222969, "url": "https://news.ycombinator.com/item?id=50222969", "by": "aryanthakor", "date": "2026-10-24", "stance": "neutral", "quote": "I've been using herdr + omp for my agent sessions", "excerpt": "I've been using herdr + omp for my agent sessions, I really like the prefix key and how easily I can edit it, does it have the similar prefix? Looks promising I will definitely give it a try. Thank you very much :)", "context": "", "use_case": "agent sessions", "switched_from": "", "story": "Show HN: Corgi, a helpful {agent, terminal} multiplexer"}, {"id": 803305122, "item_id": 50206570, "url": "https://news.ycombinator.com/item?id=50206570", "by": "arjie", "date": "2026-10-23", "stance": "neutral", "quote": "a common UI across the Claude / Codex / Oh My Pi that I use", "excerpt": "One could argue that what I really want is a tiling window manager with workspaces to swap around in, but really I want to keep my contexts separate and a command palette to swap between. e.g. https://wiki.roshangeorge.dev/w/File:Screenshot_-_Convoke_beta_build.png Obviously more interesting if I'd thought to screenshot prior to going to bed but you can have multiple tabs in a 'stream' and they can be terminal or web", "context": "", "use_case": "AI coding", "switched_from": "", "story": "A farewell to Arc browser"}, {"id": 802515905, "item_id": 50157244, "url": "https://news.ycombinator.com/item?id=50157244", "by": "Escapado", "date": "2026-10-19", "stance": "neutral", "quote": "I used Oh My Pi as a harness.", "excerpt": "Quick review: It's really fast and works really well in my initial tests. I used it to build a little Swift UI app that runs in the background and helps me OCR and translate text[1]. When I hit CMD + Shift + O it selects a portion of the screen (think of the screenshot tool), runs OCR on the portion and throws that at the google translate API and displays the result in a tooltip above / below the portion. The hotkey ", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Mercury 3: Diffusion at scale"}, {"id": 802256306, "item_id": 50141019, "url": "https://news.ycombinator.com/item?id=50141019", "by": "pimeys", "date": "2026-10-18", "stance": "recommend", "quote": "I use the omp plan mode and iterate on the text until I'm happy with it.", "excerpt": "I also used to plan with more expensive models like you did, but now I just do planning with DeepSeek too and don't see much difference. I use the omp plan mode and iterate on the text until I'm happy with it. Then clear context, give the plan to DeepSeek again for implementation. If it's mission-critical code, do a security review with GLM 5.3. Otherwise go manually through the code and ask for changes if needed. Wo", "context": "", "use_case": "plan mode for coding", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 802230385, "item_id": 50139399, "url": "https://news.ycombinator.com/item?id=50139399", "by": "nikvdp", "date": "2026-10-18", "stance": "recommend", "quote": "oh-my-pi uses a similar method and will scrape search results for you out of the box, no api key or search provider needed.", "excerpt": "What are people actually using these hosted search providers for? For personal use I've found using searxng-ai-kit [1] (disclaimer: I'm the author) or similar tools to be far more cost effective (free!). oh-my-pi uses a similar method and will scrape search results for you out of the box, no api key or search provider needed. [1]: https://github.com/nikvdp/searxng-ai-kit", "context": "", "use_case": "web search for AI agents", "switched_from": "", "story": "DeepSeek search and fetch \u2013 $0.75 per 1000 successful requests"}, {"id": 801527840, "item_id": 50095490, "url": "https://news.ycombinator.com/item?id=50095490", "by": "LaurensBER", "date": "2026-10-14", "stance": "recommend", "quote": "I've been using OMP in combination with Tern (the terminal) lately and it's an incredibly smooth operation.", "excerpt": "Interesting and I'll definitely give this a try, the client/server and plugin system do look very powerful. However, I've been using OMP in combination with Tern (the terminal) lately and it's an incredibly smooth operation. It's hard to find a decent terminal on Windows (corporate device) but this is the best I've found so far.", "context": "Windows user on a corporate device", "use_case": "", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 800071152, "item_id": 50004447, "url": "https://news.ycombinator.com/item?id=50004447", "by": "jeremyjh", "date": "2026-10-08", "stance": "recommend", "quote": "I use oh-my-pi (omp.sh) - mostly with stock settings and skills but its a \"fully loaded\" harness that you don't really have to add anything to.", "excerpt": "I use oh-my-pi (omp.sh) - mostly with stock settings and skills but its a \"fully loaded\" harness that you don't really have to add anything to. I do change a couple of things: I set it to prefer subagents, and I enable rewind. It is crazy that rewind is not a default, it saves a TON of context - if the agent goes down a crazy path that burns a lot of tokens it can rewind to an earlier checkpoint with the exploration ", "context": "", "use_case": "AI coding agent harness", "switched_from": "", "story": "Claude Haiku 5.5"}, {"id": 798923920, "item_id": 49932745, "url": "https://news.ycombinator.com/item?id=49932745", "by": "broodbucket", "date": "2026-10-02", "stance": "recommend", "quote": "As a 15+ year Arch user I used to decry people using Manjaro or the like, now I'm using omp, so I get the appeal", "excerpt": "Much like Arch you need to have a decent idea of what you want first. As a 15+ year Arch user I used to decry people using Manjaro or the like, now I'm using omp, so I get the appeal", "context": "15+ year Arch user", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798903840, "item_id": 49931490, "url": "https://news.ycombinator.com/item?id=49931490", "by": "jasonhkl_03", "date": "2026-10-02", "stance": "recommend", "quote": "I love pi simplicity but hate pi for over simplify so i use omp lolll", "excerpt": "I love pi simplicity but hate pi for over simplify so i use omp lolll", "context": "", "use_case": "AI coding harness", "switched_from": "Pi", "story": "Pi 1.0"}, {"id": 798869568, "item_id": 49929348, "url": "https://news.ycombinator.com/item?id=49929348", "by": "rpdillon", "date": "2026-10-02", "stance": "recommend", "quote": "OMP provides a significant amount of structure to the work.", "excerpt": "OMP provides a significant amount of structure to the work. That costs some tokens in terms of the system prompt. When I use these systems, I balance the return of the extra guidance in the system prompt with the cost of the context. You've talked about how much context it takes, but you haven't talked about any difference in behavior. omp offers some nice tools, like /shake, to manage context efficiently and offset ", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798868880, "item_id": 49929305, "url": "https://news.ycombinator.com/item?id=49929305", "by": "rpdillon", "date": "2026-10-02", "stance": "recommend", "quote": "I ended up choosing omp as my daily driver", "excerpt": "Yeah, saying it kills the minimalist bit misses the point of omp being a separate project. It took a minimalist-base of pi and added tools that make it usable out of the box. Hence batteries. One example here is that it added MCP extensions to make the harness usable with MCP servers long before Pi figured out that that was the right move. A lot can be said about development of MCP over that timeframe, but I apprecia", "context": "Long-time Linux and vanilla Emacs user", "use_case": "AI coding harness", "switched_from": "Pi", "story": "Pi 1.0"}, {"id": 798868353, "item_id": 49929272, "url": "https://news.ycombinator.com/item?id=49929272", "by": "nixpulvis", "date": "2026-10-02", "stance": "neutral", "quote": "I use both `pi` and `omp`", "excerpt": "A coworker was trying to tell me that models perform better in their own agent harnesses. I use both `pi` and `omp` and I'm somewhat skeptical. I understand that the tool calls might be slightly different. But really how much impact on the model itself does the harness have?", "context": "", "use_case": "agent harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798861888, "item_id": 49928868, "url": "https://news.ycombinator.com/item?id=49928868", "by": "gchamonlive", "date": "2026-10-02", "stance": "neutral", "quote": "I use oh-my-pi, not sure how it compares", "excerpt": "I use oh-my-pi, not sure how it compares, but I say someone praising antigravity for being a good harness(1), and for the love of good people settle for such low standards of user experience it's almost pitiful. (1) https://news.ycombinator.com/item?id=49913854", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798854833, "item_id": 49928427, "url": "https://news.ycombinator.com/item?id=49928427", "by": "orsorna", "date": "2026-10-02", "stance": "compare", "quote": "Do users of OMP find all of its additions useful?", "excerpt": "I am wondering what use people get out of this. What I like about pi is that, using proprietary harnesses to begin with, I already somewhat know what kind of features I use and features I wish I could use. Asking pi to self modify itself (with a smart enough model) is a one time cost for a fully vetted feature that I personally approve. Do users of OMP find all of its additions useful?", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "Pi 1.0"}, {"id": 798853984, "item_id": 49928374, "url": "https://news.ycombinator.com/item?id=49928374", "by": "CharlesW", "date": "2026-10-01", "stance": "recommend", "quote": "I recommend the batteries-included Pi distribution oh-my-pi (omp.sh) over Pi and OpenCode.", "excerpt": "As someone similarly invested in Claude Code and initially reluctant to try other harnesses, I recommend the batteries-included Pi distribution oh-my-pi (omp.sh) over Pi and OpenCode. I switch between Claude Code, omp, and Codex regularly, and it feels fine.", "context": "heavily invested in Claude Code", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798853040, "item_id": 49928315, "url": "https://news.ycombinator.com/item?id=49928315", "by": "CharlesW", "date": "2026-10-01", "stance": "recommend", "quote": "I can strongly recommend omp.sh if you just want to be up and running without hassle and a far better (for me) developer experience.", "excerpt": "That the \"batteries included\" bit, but it's not \"burning tokens\" in the sense that using Pi with the same batteries would (of course) use just as many. I can strongly recommend omp.sh if you just want to be up and running without hassle and a far better (for me) developer experience.", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Pi 1.0"}, {"id": 798837841, "item_id": 49927365, "url": "https://news.ycombinator.com/item?id=49927365", "by": "razster", "date": "2026-10-01", "stance": "warn", "quote": "I've tried Oh-my-pi but it's too heavy for me and eats up 6% of my context on start.", "excerpt": "I use my Pi to organize my graphic design assets as well as update documentation for client projects. It also is my daily driver for a story I'm working on and thanks to a ton of great skills created it has become more powerful. Local model is Qwen3.6 35B-A3B and Qwen3.8 27B UD. I love the compressing function, keeps on pushing. I've tried Oh-my-pi but it's too heavy for me and eats up 6% of my context on start. Also", "context": "", "use_case": "", "switched_from": "", "story": "Pi 1.0"}, {"id": 798391056, "item_id": 49899441, "url": "https://news.ycombinator.com/item?id=49899441", "by": "pimeys", "date": "2026-09-29", "stance": "neutral", "quote": "I have 7 agents in omp running 24/7 every day.", "excerpt": "How on earth you can do 100 dollars in 2-3 days with DeepSeek? I have 7 agents in omp running 24/7 every day. I use maybe 10-15 dollars a day. A rarely see a session going over 2 dollars. My maximum is maybe 3.5 dollars and that session took three days. What harness you are using?", "context": "", "use_case": "running coding agents", "switched_from": "", "story": "GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price"}, {"id": 798390704, "item_id": 49899419, "url": "https://news.ycombinator.com/item?id=49899419", "by": "taylorfinley", "date": "2026-09-29", "stance": "recommend", "quote": "I cannot recommend Oh My Pi enough.", "excerpt": "Maybe CC does something that breaks the cache? I cannot recommend Oh My Pi enough. Every default is galaxy brained, and it plays incredibly well with deepseek flash 4.1. My favorite coding harness rn for sure.", "context": "", "use_case": "coding harness", "switched_from": "", "story": "GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price"}, {"id": 797642432, "item_id": 49852652, "url": "https://news.ycombinator.com/item?id=49852652", "by": "spamizbad", "date": "2026-09-26", "stance": "recommend", "quote": "I greatly enjoy using the plan mode that comes baked into omp (oh my pi)", "excerpt": "It's been a while (well, 8 months) since I used Claude Code, but what I can say is that I greatly enjoy using the plan mode that comes baked into omp (oh my pi), especially for tasks around firmware development. I recognize that has a very different workflow than webapps though.", "context": "firmware developer", "use_case": "firmware development planning", "switched_from": "Claude Code", "story": "Plan mode is dead"}, {"id": 797473137, "item_id": 49842071, "url": "https://news.ycombinator.com/item?id=49842071", "by": "akmarinov", "date": "2026-09-25", "stance": "neutral", "quote": "for coding i use OMP directly with Deepseek", "excerpt": "I'm using the Deepseek harness with the deepseek computer use and deepseek browser integration, it's somewhat janky but it gets the job done. Also for coding i use OMP directly with Deepseek. It's been rather awesome because even maxed out $20 gets you very far in between Codex resets. I also tried developing a small game where I have Astra use as many Deepseek workers and it needs and it only does the 3D objects and", "context": "", "use_case": "coding harness", "switched_from": "", "story": "Best LLM for every budget, updated daily"}, {"id": 796936433, "item_id": 49808527, "url": "https://news.ycombinator.com/item?id=49808527", "by": "jeremyjh", "date": "2026-09-22", "stance": "recommend", "quote": "In OMP I have Sol high as the orchestrator, Sol max as Planner & Reviewer, Luna max as task/coder. Very good setup.", "excerpt": "You can get a lot of Codex usage out of that same sub on top of ChatGPT usage. Its a really good value and you can use that sub in any harness. In OMP I have Sol high as the orchestrator, Sol max as Planner & Reviewer, Luna max as task/coder. Very good setup. I'm on pro now and there are weekends when I use half a week's usage but I'll have 5 or 6 sessions going at once for many hours each day.", "context": "on Pro plan", "use_case": "multi-agent coding harness", "switched_from": "", "story": "GPT-6 Sol and Luna"}, {"id": 796919072, "item_id": 49807442, "url": "https://news.ycombinator.com/item?id=49807442", "by": "glub", "date": "2026-09-22", "stance": "recommend", "quote": "I have been using oh-my-pi with 3 claude subs for the past few months without any issues.", "excerpt": "Not natively, as it's still a ToS violation and adding that in pi would go against pi principles, but there are many plugins/proxies that make it work. oh-my-pi supports it natively (again, still a ToS violation), by impersonating claude code's fingerprints. I have been using oh-my-pi with 3 claude subs for the past few months without any issues. Even native server-side OAI/ANT compaction works out of the box.", "context": "", "use_case": "using Claude subscriptions in a harness", "switched_from": "", "story": "GPT-6 Sol and Luna"}, {"id": 796289680, "item_id": 49768105, "url": "https://news.ycombinator.com/item?id=49768105", "by": "desipenguin", "date": "2026-09-19", "stance": "recommend", "quote": "I use oh-my-pi (and opencode, and toad) but use (client paid) Opus 5 with these harnesses.", "excerpt": "Genuine question: Does the Agent harness matter ? I mean why use Claude CLI ? (As I mentioned elsewhere) I use oh-my-pi (and opencode, and toad) but use (client paid) Opus 5 with these harnesses. All three of these understand AGENTS.md", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Claude Code now reads AGENTS.md if there is no Claude.md"}, {"id": 796289072, "item_id": 49768067, "url": "https://news.ycombinator.com/item?id=49768067", "by": "desipenguin", "date": "2026-09-19", "stance": "recommend", "quote": "I also use the same. But I use Anthropic models (Opus 5) with it.", "excerpt": "Do you mean oh-my-pi (omp) (https://github.com/can1357/oh-my-pi/) I also use the same. But I use Anthropic models (Opus 5) with it.", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "Claude Code now reads AGENTS.md if there is no Claude.md"}, {"id": 795685539, "item_id": 49730346, "url": "https://news.ycombinator.com/item?id=49730346", "by": "pimeys", "date": "2026-09-16", "stance": "neutral", "quote": "My harness: https://omp.sh/", "excerpt": "My experience is complete opposite from yours. The V4 Flash was already quite good, but V4.1 is really very close to SOTA in my books. I've eval'd these models for weeks against Gemini 3.8, Kimi K3 and Opus 5, and DeepSeek absolutely wins these evals. It's really crazy. We price per token our customers. If Kimi was about 30% of the price of Opus 5 for the same quality, DeepSeek is 1/10th of a price of Kimi K3. We've ", "context": "", "use_case": "agent harness", "switched_from": "", "story": "DeepSeek v4.1 Flash Is Now Our Best Hacking Model"}, {"id": 794778722, "item_id": 49673670, "url": "https://news.ycombinator.com/item?id=49673670", "by": "tiluha", "date": "2026-09-12", "stance": "neutral", "quote": "Using omp as harness.", "excerpt": "My current vibe/playing around setup is deepseek v4.1 flash with opus/sol as advisor agent, dsv4.1flash agents for self review and final review with opus/sol. I love how fast and cheap deepseek is. Using omp as harness. Privacy policy is not great with deepseek api, but you are always just trusting their word with any hosted llm and in theory i could at least self host the models i use from deepseek.", "context": "", "use_case": "agent harness", "switched_from": "", "story": "Ask HN: What default model do you use and why?"}, {"id": 793699377, "item_id": 49606211, "url": "https://news.ycombinator.com/item?id=49606211", "by": "alvins82", "date": "2026-09-08", "stance": "neutral", "quote": "I pair it with OMP via https://github.com/alvins82/omp-openchamber-server/.", "excerpt": "Btw in my quest for a good desktop codex/claude like app - https://github.com/openchamber/openchamber - this seems to be front-runner. I pair it with OMP via https://github.com/alvins82/omp-openchamber-server/. I wanted a powerful GUI+harness setup for open models so I could use/test as they came out.", "context": "", "use_case": "harness for open models", "switched_from": "", "story": "I tested 10 model/harness combinations on the same Three.js task"}, {"id": 793257537, "item_id": 49578596, "url": "https://news.ycombinator.com/item?id=49578596", "by": "latchkey", "date": "2026-09-05", "stance": "recommend", "quote": "Using omp helps a lot too, the fact that it uses an LSP under the covers is a game changer.", "excerpt": "I'm addicted. This is the most fun I've had coding in decades. I feel like I've got superpowers and the limit is just how quickly I can prompt out my ideas. Those 500m are for coding. Started with $20 and now I'm on the $200/mo codex plan. Resets are crazy... free tokens, why not take advantage of all that VC money? Using omp helps a lot too, the fact that it uses an LSP under the covers is a game changer. I'm buildi", "context": "", "use_case": "AI coding harness", "switched_from": "", "story": "Ask HN: Who is using MCP in production?"}, {"id": 792722176, "item_id": 49545136, "url": "https://news.ycombinator.com/item?id=49545136", "by": "benw214", "date": "2026-09-03", "stance": "recommend", "quote": "OMP i would say comes out on this prompt the best out of all the harnesses i've tried.", "excerpt": "I'm not sure i agree with these. I've been testing a simple prompt with different model and harness combinations to work out which one produces best results. I do this in '/goal' mode. Prompt - \"Build a single-page Three.js sci-fi hangar with hovering drones, animated warning lights, emissive runway strips, and subtle volumetric-style fog planes. Include drone formation toggle and cinematic camera path. Output one se", "context": "testing harness and model combinations", "use_case": "coding agent harness", "switched_from": "", "story": "Show HN: FrontierHarness Eval \u2013 9 harness, same model, cost per pass varies 17x"}, {"id": 790522193, "item_id": 49407637, "url": "https://news.ycombinator.com/item?id=49407637", "by": "pimeys", "date": "2026-08-23", "stance": "recommend", "quote": "Except a month ago I started testing Kimi K3 and omp and I never went back.", "excerpt": "I've been using all the SOTA models a lot at work, like serious amount of tokens. It's been really rare that I stick with one model and harness for too long... Except a month ago I started testing Kimi K3 and omp and I never went back. Something with this combo works really well for Rust dev. The model doesn't really annoy me at all and I have not switched to Opus or SOL. And the monthly token bill is much lower...", "context": "heavy work user of SOTA models", "use_case": "coding agent harness", "switched_from": "", "story": "A week of using Codex more than Claude"}, {"id": 790256516, "item_id": 49391032, "url": "https://news.ycombinator.com/item?id=49391032", "by": "dannyw", "date": "2026-08-21", "stance": "recommend", "quote": "Try something like Pi or OMP. Models perform better in these harnesses than Claude Code", "excerpt": "Are you using an API, or running locally? If so, are you running with a quant, or other 'optimisations'? I've been using it via openrouter pretty heavily as my daily driver for the past week and loving it, have never experienced incoherent rubbish even at 500k+ contexts (that's usually way higher than I'd typically compact at), and tool calling reliability is better than Opus 5 in the Claude Code harness. Modern Anth", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "DeepSeek-v4-flash-vision-exp"}, {"id": 789973904, "item_id": 49373369, "url": "https://news.ycombinator.com/item?id=49373369", "by": "rurban", "date": "2026-08-20", "stance": "recommend", "quote": "Behaves better than opencode, and I love their /collab remote control.", "excerpt": "This sound exactly like the new DeepSeek Harness cli. But I switched to oh-my-pi and crof.ai recently (cheap deepseek). Behaves better than opencode, and I love their /collab remote control.", "context": "", "use_case": "coding agent harness", "switched_from": "opencode", "story": "fx :Tiny, open, native coding agent."}, {"id": 789969329, "item_id": 49373083, "url": "https://news.ycombinator.com/item?id=49373083", "by": "DannyBee", "date": "2026-08-20", "stance": "recommend", "quote": "falling well behind the other one i use (OMP) in the past 6 months in usability", "excerpt": "So many developer-hostile things lately. Just a few days ago they turned on an experiment the forces claude code to use bash over standard tools in auto mode. From the system prompt directly, new as of August 18th: Do your work through the Bash tool wherever it can accomplish the job: read files with cat, head, or sed -n, search with grep and find, and make file changes with sed, heredocs, or short scripts, rather th", "context": "", "use_case": "coding agent harness", "switched_from": "Claude Code", "story": "Anthropic Refuses to Support Agents.md"}, {"id": 789902850, "item_id": 49368928, "url": "https://news.ycombinator.com/item?id=49368928", "by": "andai", "date": "2026-08-19", "stance": "compare", "quote": "250k LoC and 1.5M LoC respectively. Sheesh. How's that for minimalism...", "excerpt": "The sub only works in Claude Code itself (unless they changed that rule again?), and all the Chinese models are trained on Claude Code (a bunch of them don't even work in Codex). I use a custom harness, but I constantly hear good things about Pi. EDIT: Checked out Pi and Oh-My-Pi. 250k LoC and 1.5M LoC respectively. Sheesh. How's that for minimalism... Mine's 50. (Lines.) https://gist.github.com/a-n-d-a-i/bd50aaa4bdb", "context": "uses a custom harness", "use_case": "", "switched_from": "", "story": "Anthropic Refuses to Support Agents.md"}, {"id": 789266305, "item_id": 49329144, "url": "https://news.ycombinator.com/item?id=49329144", "by": "phiagent", "date": "2026-08-17", "stance": "neutral", "quote": "omp.sh has lots of interesting features. I used it for a while, but found it a bit complicated", "excerpt": "Yeah, omp.sh has lots of interesting features. I used it for a while, but found it a bit complicated, so I ended up switching back to Pi. I stick with Pi because I believe a core tool ought to stay simple. I\u2019ve built my own coding agent using Pi and Go.", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Ask HN: Which plugins or extensions do you most enjoy using with the Pi agent?"}, {"id": 788984993, "item_id": 49311562, "url": "https://news.ycombinator.com/item?id=49311562", "by": "217", "date": "2026-08-15", "stance": "recommend", "quote": "omp.sh if youre into figuring out and tinkering with the best tools available", "excerpt": "Just make skills for things you're doing repeatedly / often correcting the model on. If you are just starting i highly recommend trying out codex desktop app (it's good and just works, especially if you are on the plan that has 5.6 sol) or omp.sh if youre into figuring out and tinkering with the best tools available", "context": "", "use_case": "tinkering with agent tooling", "switched_from": "", "story": "Ask HN: What agent skills are necessary?"}, {"id": 788761905, "item_id": 49297619, "url": "https://news.ycombinator.com/item?id=49297619", "by": "mattkevan", "date": "2026-08-14", "stance": "recommend", "quote": "I like the Oh My Pi harness, but the site's insufferable", "excerpt": "It's really annoying. I've had to write a CLAUDE.md file that specifically bans particular phrases and tries to keep narrative out of comments. Also the I have ADHD skill [1] helps to force Opus to get to the point. It's also the case when using Claude Design - it loves to fill the UI with little labels that describe how everything works. I think it's been trained on both UI microcopy and functional annotations and c", "context": "", "use_case": "coding harness", "switched_from": "", "story": "Why does Opus 5 feel worse to work with?"}, {"id": 788760336, "item_id": 49297521, "url": "https://news.ycombinator.com/item?id=49297521", "by": "stpedgwdgfhgdd", "date": "2026-08-14", "stance": "recommend", "quote": "Just switched to oh-my-pi, it has gotten pretty good.", "excerpt": "Just switched to oh-my-pi, it has gotten pretty good. For example the web-search is nice. Subagents, if you want to\u2026 Cmux, Sol and omp are my tools for now. CC is just too expensive for usage-based pricing.", "context": "", "use_case": "coding harness", "switched_from": "Claude Code", "story": "Why does Opus 5 feel worse to work with?"}, {"id": 787732000, "item_id": 49233250, "url": "https://news.ycombinator.com/item?id=49233250", "by": "latchkey", "date": "2026-08-09", "stance": "recommend", "quote": "i just switched to it from codex... and i feel like i just leveled up. what sells it for me is that it uses LSP's to edit the files.", "excerpt": "i was avoiding the harness wars, but a friend recommended it and i just switched to it from codex... and i feel like i just leveled up. what sells it for me is that it uses LSP's to edit the files. so much faster and more accurate edits. use it with ACP through Zed and you get rid of all the TUI graphics nonsense...", "context": "", "use_case": "coding agent harness", "switched_from": "Codex", "story": "I Wanted to Own the Harness. Then Codex Desktop Won"}, {"id": 787220130, "item_id": 49201258, "url": "https://news.ycombinator.com/item?id=49201258", "by": "eli", "date": "2026-08-06", "stance": "recommend", "quote": "OpenCode or oh-my-pi might make more sense if you just want a batteries-included agent.", "excerpt": "I use https://pi.dev/ which works fine out of the box but is fairly minimal and intended to be customized. There are many extensions. OpenCode or oh-my-pi might make more sense if you just want a batteries-included agent. You can also make Claude Code work with other models without too much work, but I think that's asking for headaches.", "context": "", "use_case": "CLI coding agent", "switched_from": "", "story": "Qwen3.8 Max now ranked as the best overall model by agentic index"}, {"id": 784478160, "item_id": 49029885, "url": "https://news.ycombinator.com/item?id=49029885", "by": "greggh", "date": "2026-07-23", "stance": "recommend", "quote": "Stock OMP. When you add a provider and then /model you can choose a model and choose what roles it will take.", "excerpt": "Stock OMP. When you add a provider and then /model you can choose a model and choose what roles it will take. Do that for each role it has. Then I use /plan when I need to be planning and not writing to any files. Take the advice the UI gives you where it tells you to add the word orchestrate into your prompts when you want to make sure it uses its todo/tasks and sub-agents.", "context": "", "use_case": "agentic coding with planning", "switched_from": "", "story": "Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA"}, {"id": 783830192, "item_id": 48989387, "url": "https://news.ycombinator.com/item?id=48989387", "by": "217", "date": "2026-07-21", "stance": "recommend", "quote": "most advanced, configurable and capable harness of all time", "excerpt": "https://github.com/can1357/oh-my-pi https://stencil.so/blog/prewalk most advanced, configurable and capable harness of all time, and i use both 200$ subs with it. Has a huge learning curve though so if you're not into that i'd go with whatever you already use the most", "context": "pays for two $200 subscriptions", "use_case": "coding agent harness", "switched_from": "", "story": "Ask HN: Claude Code or Codex?"}, {"id": 781580480, "item_id": 48848780, "url": "https://news.ycombinator.com/item?id=48848780", "by": "roger_", "date": "2026-07-09", "stance": "recommend", "quote": "Currently using Oh My Pi (https://github.com/can1357/oh-my-pi) and appreciate the batteries included approach.", "excerpt": "Currently using Oh My Pi (https://github.com/can1357/oh-my-pi) and appreciate the batteries included approach. From my limited time using pure Pi, I found quite a few of the plugins lacking and had no desire to upgrade/fix and maintain them myself. I know others feel differently though. I like the idea of keep Pi minimal but having \u201cofficial\u201d, high quality optional plugins to make it more usable.", "context": "", "use_case": "coding agent config", "switched_from": "Pi", "story": "Opinionated and Easy Pi.dev Configuration"}, {"id": 777826080, "item_id": 48614130, "url": "https://news.ycombinator.com/item?id=48614130", "by": "Bnjoroge", "date": "2026-06-20", "stance": "recommend", "quote": "I use OMP. Was on Pi but switched over because I like OMP's subagents implementation.", "excerpt": "I use OMP. Was on Pi but switched over because I like OMP's subagents implementation. Use 5.5/4.8 for coming up with a very detailed plan, ds4pro/mimo for implementation of easy features, 5.5/4.6 for larger/harder features, and 5.5/4.6 for a review. I use them in onorca.dev and sometimes use Zed as well.", "context": "", "use_case": "AI coding agent", "switched_from": "Pi", "story": "Ask HN: Do you use Claude Code, Codex, or something else?"}]}