# Pi coding agent

892 Hacker News commenters recommended Pi coding agent and 198 criticized it, November 2025 to November 2026. It was recommended in 12 of 13 months; owners who said how long report a median of 0.25 years (51 owners). It meets the Hacker Approved badge rule.

**Verdict:** A minimal coding agent harness that works with many models, uses fewer tokens and is easy to extend, though basics like subagents are missing and it has no sandboxing.

## Why people recommend it

- Minimal and highly customizable (9 comments): "It's a pleasure to work with, and so far it's been my experience that if pi doesn't behave the way I want, it's pretty easy to modify to suit my needs." ([source](https://news.ycombinator.com/item?id=48804958)) "Using pi with tons of custom extensions, that to be fair I developed since making the jump off of codex and claude about 12 weeks ago." ([source](https://news.ycombinator.com/item?id=49793444)) "plugins for executables suck compared to "just create a .ts file in this directory and edit/test it in a few seconds until satisfied"" ([source](https://news.ycombinator.com/item?id=49930280))
- Small prompt saves tokens (6 comments): "Claude code includes a large system prompt with every request while tool like pi does not. This could save tokens resulting in lower costs." ([source](https://news.ycombinator.com/item?id=47153168)) "It would be fairer to compare K3 using the same generic harness, such as a Pi setup with some sane extensions for token optimisation." ([source](https://news.ycombinator.com/item?id=49045989)) "Pi’s minimal system prompt avoids contradiction between what the harness writer thinks is best and what the user thinks is best." ([source](https://news.ycombinator.com/item?id=49623397))
- Works well across many models (6 comments): "I switched to Pi and chinese open source models, and models that _I know_ are less capable than sonnet outperform my sonnet + claude code stack at work." ([source](https://news.ycombinator.com/item?id=48590561)) "Since models are being released every few weeks this is interesting to compare models without having to switch to a different harness." ([source](https://news.ycombinator.com/item?id=47155345)) "As models get better they will get better at using any harness, it comes down to how well the harness is actually engineered." ([source](https://news.ycombinator.com/item?id=48242300))
- Easy to write extensions (4 comments): "It seems pretty well done: when I added permission requests to the `bash` tool, the "Are you sure y/N" requests started appearing just like they were native to Emacs." ([source](https://news.ycombinator.com/item?id=47154047)) "The most interesting thing for me is that I built an extension for Pi that has it recognize when it does not know how to do something" ([source](https://news.ycombinator.com/item?id=47024661)) "plugins for executables suck compared to "just create a .ts file in this directory and edit/test it in a few seconds until satisfied"" ([source](https://news.ycombinator.com/item?id=49930280))

## What people warn about

- Requires lots of tinkering (3 comments): "if you like spending your time tweaking and customizing the tool itself rather than doing the work, it will be the perfect match" ([source](https://news.ycombinator.com/item?id=50326730)) "Meanwhile, you wasted those tokens and time that could have been spent building actual, useful software instead of hobby tinkering your harness." ([source](https://news.ycombinator.com/item?id=48270360)) "I don't want to spend 2 hours prompting, configuring and fixing features that I need which are standard in every other harness." ([source](https://news.ycombinator.com/item?id=49177746))
- Missing built-in basic features (3 comments): "I don't want to spend 2 hours prompting, configuring and fixing features that I need which are standard in every other harness." ([source](https://news.ycombinator.com/item?id=49177746)) "having said that, i find it annoying and not empowering that basic things like subagents and web search aren't built in" ([source](https://news.ycombinator.com/item?id=49926430)) "Pi doesn't have a UI like Claude Desktop. It also doesn't work with the Claude subscription, only API key and pricing." ([source](https://news.ycombinator.com/item?id=49409535))
- No sandboxing or security (2 comments): "But pi doesn't have security. Pi is full yolo, it's on the user to run it in an environment that minimizes the blast radius" ([source](https://news.ycombinator.com/item?id=49907798)) "it went out of the app directory it started in, and made code edits to a sibling directory in a different git project repo" ([source](https://news.ycombinator.com/item?id=49177640))
- Context overflow problems (2 comments): "struggling with local AI in Pi to even get small (500-1000 lines of code) programs written without overrunning a 220k context" ([source](https://news.ycombinator.com/item?id=48946611)) "It choked at some point and stopped responding (probably a context overflow before pi-coding-agent could compact it)." ([source](https://news.ycombinator.com/item?id=48014659))

## Alternatives hackers recommend

- [omp (oh-my-pi)](https://hackerapproved.com/p/omp.md): 80 recommend
- [ForgeCode](https://hackerapproved.com/p/forge.md): 39 recommend
- [Reasonix](https://hackerapproved.com/p/reasonix.md): 26 recommend
- [DSH](https://hackerapproved.com/p/dsh.md): 11 recommend
- [smol](https://hackerapproved.com/p/smol.md): 9 recommend

## Cite as

Hacker Approved, "Pi coding agent", https://hackerapproved.com/p/pi/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).