# ForgeCode

39 Hacker News commenters recommended ForgeCode and 12 criticized it, November 2025 to November 2026. It was recommended in 10 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Free coding agent harness that people credit with reliable tool calling for small local models and strong Terminal Bench results; critics question its benchmark claims.

## Why people recommend it

- Tool calling for small models (3 comments): "Forge just increases mechanical reliability of tool-calling, it shouldn't disrupt your workflow-level layers much." ([source](https://news.ycombinator.com/item?id=48207188)) "Forge provides a more hardened suite of guardrails that are effectively necessary for small models." ([source](https://news.ycombinator.com/item?id=48203713)) "It can bring 8b models up to Opus-like performance at tool calling." ([source](https://news.ycombinator.com/item?id=48524199))
- Strong Terminal Bench results (2 comments): "ForgeCode appears to be pretty good (currently, #1 on Terminal Bench 2.0" ([source](https://news.ycombinator.com/item?id=47879014)) "has better results than Open Code, at least according to Terminal Bench" ([source](https://news.ycombinator.com/item?id=48005639))
- Preferred over Claude Code (2 comments): "I really liked forgecode the best, and it feels better than claude code" ([source](https://news.ycombinator.com/item?id=48196684)) "Firstly, forgecode is a much better harness than Cloude code (IMHO)." ([source](https://news.ycombinator.com/item?id=48503629))
- Structured planning and step-by-step execution (2 comments): "Forge brings rigorous, adversarial architectural planning and reliable step-by-step execution to AI-assisted development." ([source](https://news.ycombinator.com/item?id=47135289)) "keeps a decision log and a dependency-aware task graph so changes remain explainable and resumable" ([source](https://news.ycombinator.com/item?id=47309023))

## What people warn about

- Accused of gaming benchmarks (2 comments): "I cannot believe they didn't do something to influence the benchmarks" ([source](https://news.ycombinator.com/item?id=47881871)) "ForgeCode for instance has been cheating in the past" ([source](https://news.ycombinator.com/item?id=48005766))
- Stalls when model replies with text (2 comments): "But at the end of the day, if the model keeps responding with text, there's nothing forge can do." ([source](https://news.ycombinator.com/item?id=48203924)) "if the model keeps responding with text, there's nothing forge can do" ([source](https://news.ycombinator.com/item?id=48205670))

## Alternatives hackers recommend

- [Pi coding agent](https://hackerapproved.com/p/pi.md): 892 recommend
- [omp (oh-my-pi)](https://hackerapproved.com/p/omp.md): 80 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, "ForgeCode", https://hackerapproved.com/p/forge/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).