Crush by Charmbracelet
Hacker ApprovedRanked 9 of 18 AI coding agents by recommenders
Free open-source terminal coding agent that works with any AI provider, praised for its polished interface and language server support. Main caveat: docs are thin for local models.
Free open-source terminal coding agent that works with any AI provider, praised for its polished interface and language server support. Main caveat: docs are thin for local models.
47 people gave an opinion
- recommend it
- 39
- criticize it
- 8
- approval share
- 83%
Recommended in 12 of 13 months
- November 2025: recommended
- December 2025: recommended
- January 2026: recommended
- February 2026: recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it carries the badge
- 39 people recommend itNeeds 10 or more, met.
- 4.9 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 12 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Polished, pleasant interface and layout
“lots of effort put in to layouts, color, and really pushing the boundaries of what a TUI can be”
Comment from 23 Apr 2026 in “SpaceX and Cursor partnership. Right to acquire Cursor later this year” “It has the most user-friendly UI of them all, a really usable scroll view.”
Comment from 22 Jul 2026 in “You only need the frontier model for one single edit” “by far the prettiest and easiest to reason about what is going on”
Comment from 15 Aug 2026 in “Qwen 3.8 27B is out: open weights, best local dense model yet” - Comment from 3 Feb 2026 in “Claude Code is suddenly everywhere inside Microsoft”
- Comment from 12 Nov 2025 in “Testing out Crush, a TUI based coding agent”
Solid results with open models
“I've been using charm's Crush with GLM for several months and it's been working great.”
Comment from 12 Jun 2026 in “Kimi K2.7-Code: open-source coding model with better token efficiency” “Crush+GLM is pretty good at following plans done by Claude or Codex”
Comment from 12 Jan 2026 in “Remote Code Execution in OpenCode, update now” “I’ve been using Crush with Openrouter and have good success lately”
Comment from 1 Jul 2026 in “ZCode: Claude Code from the Makers of GLM” - Comment from 24 Apr 2026 in “I Cancelled Claude: Token Issues, Declining Quality, and Poor Support”
- Comment from 14 Aug 2026 in “Qwen 3.8 27B is out: open weights, best local dense model yet”
Works with any model provider
“If you want open source and want to target something over an API "crush" https://github.com/charmbracelet/crush is excellent”
Comment from 1 Nov 2025 in “Claude Code can debug low-level cryptography” “API-neutral tools like Charm Crush which can be used with any AI provider with API keys, and work reasonably well”
Comment from 21 Jan 2026 in “Vibecoding #2” “They maintain a list of supported providers so most of the time no configuration is needed.”
Comment from 7 Jul 2026 in “Ask HN: What is your AI harness that lets you switch LLM models easily?” - Comment from 10 Dec 2025 in “Getting a Gemini API key is an exercise in frustration”
Lightweight, low resource use
“Written in Go, compiled binary, respects XDG (e.g. uses ~/.config/crush to store config), etc.”
Comment from 6 Aug 2026 in “Pi's Minimalism Is Its Advantage” “highly recommended if you want something which can work on low constrained devices”
Comment from 21 Mar 2026 in “OpenCode – The open source AI coding agent” “compared to node-based harnesses, uses 1/5th the RAM”
Comment from 24 May 2026 in “DeepSeek reasonix, engineered for low token cost during long running sessions”
What people warn about
Each mark is one comment
Docs are thin or outdated
“Pretty straightforward but agree that their docs could be better for local LLM setups in particular.”
Comment from 24 Apr 2026 in “I Cancelled Claude: Token Issues, Declining Quality, and Poor Support” “the docs are all over the place on linking into arbitrary APIs”
Comment from 24 Apr 2026 in “I Cancelled Claude: Token Issues, Declining Quality, and Poor Support” “crush has outdated docs and abstract credit based pricing, both big red flags for me”
Comment from 19 Oct 2026 in “OpenCode has a security breach last that impacted 1.4% of console users” Fewer features than rivals
“less feature rich than Claude Code, but with a smaller system prompt and better built-in LSP support”
Comment from 15 Jun 2026 in “Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?” “10/10 pretty but lacking in features I need.”
Comment from 21 Mar 2026 in “OpenCode – The open source AI coding agent” Prefer other harnesses like Pi
“ultimately if I had a nitpick, I would look back over at pi, and the grass looked greener.”
Comment from 18 Jun 2026 in “Migrate from OpenClaw” “It's so-so though I prefer Pi myself.”
Comment from 24 May 2026 in “DeepSeek reasonix, engineered for low token cost during long running sessions” Proprietary and credit pricing concerns
“crush has outdated docs and abstract credit based pricing, both big red flags for me”
Comment from 19 Oct 2026 in “OpenCode has a security breach last that impacted 1.4% of console users” “Crush is notably and disappointingly proprietary”
Comment from 12 Nov 2025 in “Testing out Crush, a TUI based coding agent”
Other AI coding agents people recommend
The four most recommended other AI coding agents, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Claude Code | 4,522 | 3,788 | 54% | 13 | 6 mo417 owners | — |
| OpenAI Codex | 2,474 | 1,289 | 66% | 13 | 3 mo165 owners | — |
| OpenCode | 1,060 | 572 | 65% | 13 | 2 mo41 owners | — |
| Gemini CLI | 172 | 321 | 35% | 12 | 3 mo9 owners | — |
| Crush by CharmbraceletThis page | 39 | 8 | 83% | 12 | — | Approved |
What people said
40 of the 78 comments that mention it, newest first. Each links to the original comment.
“crush has outdated docs and abstract credit based pricing, both big red flags for me”
“They work well, except that the agent will often simply call wait() immediately after”
“It's hard to find good harnesses that don't pull in mounds of Javascript”
“by far the prettiest and easiest to reason about what is going on”
“Use 3.6 27b as a daily driver for months with charmbracelet crush”
“I think we Linux users have got this without Sams help.”
“Written in Go, compiled binary, respects XDG (e.g. uses ~/.config/crush to store config), etc.”
“It has the most user-friendly UI of them all, a really usable scroll view.”
“There are several (pi, opencode, crush, etc) and then clone that repo”
“I will let it access *.openrouter.ai (LLM providers) but not *.charm.land”
“They maintain a list of supported providers so most of the time no configuration is needed.”
“I’ve been using Crush with Openrouter and have good success lately”
“ultimately if I had a nitpick, I would look back over at pi, and the grass looked greener.”
“I use crush as coding agent but removed the blocked commands”
“less feature rich than Claude Code, but with a smaller system prompt and better built-in LSP support”
“I've been using charm's Crush with GLM for several months and it's been working great.”
“compared to node-based harnesses, uses 1/5th the RAM”
“It's so-so though I prefer Pi myself.”
“don't think any of the current harnesses, Claude Code, Pi, Crush, OpenCode etc do that”
“Crush (from the charm-bracelet folks) is particularly nice when starting out”
“Pretty straightforward but agree that their docs could be better for local LLM setups in particular.”
“the docs are all over the place on linking into arbitrary APIs”
“Having some success with Qwen+Crush and have been more productive.”
“lots of effort put in to layouts, color, and really pushing the boundaries of what a TUI can be”
“I've used it to make up a couple of internal projects without any issues.”
“a standalone go binary by the original developer of OpenCode”
“10/10 pretty but lacking in features I need.”
“highly recommended if you want something which can work on low constrained devices”
“we did a lot of work on improving the experience, both on UX, performance, and the actual reliability of the agent itself”
“Crush is their "opencode"-ish app and is my go to when not using CC or CopilotCLI directly.”
“The best UX is done by Crush. and they nail the experience”
“API-neutral tools like Charm Crush which can be used with any AI provider with API keys, and work reasonably well”
“I use it heavily as a "I don't want to spend my anthropic credits" day-to-day model (mostly using Crush)”
“Crush+GLM is pretty good at following plans done by Claude or Codex”
“I find it handy in that I can pull it into any IDE and it works the same.”
“I think the magic sauce is the fact that they have language server support”
“They could always just use OpenCoder, Crush or Goose with OpenRouter”
“crush looks nice and I like the wacky vibe of charm”
“Crush is notably and disappointingly proprietary”
“If you want open source and want to target something over an API "crush" https://github.com/charmbracelet/crush is excellent”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.