Gas Town
The most recommended of 26 agent orchestrators
A free tool for coordinating many AI coding agents, valued for its orchestration ideas, though critics say it is expensive to run, dangerous and fully vibe-coded.
A free tool for coordinating many AI coding agents, valued for its orchestration ideas, though critics say it is expensive to run, dangerous and fully vibe-coded.
307 people gave an opinion
- recommend it
- 95
- criticize it
- 212
- approval share
- 31%
Recommended in 10 of 13 months
- November 2025: not recommended
- December 2025: not 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)
Owned for 1 month
Median of the 3 owners who said how long they have had it.
Each square is one year.
Why it doesn’t carry the badge
- 95 people recommend itNeeds 10 or more, met.
- 0.4 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 10 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
Orchestration and validator design
“the idea of having orchestrator threads to manage your work and keep tabs on it, having validators to validate the other work etc. were generally the right shape”
Comment from 25 Aug 2026 in “Fences, Not Sandboxes” “it does seem to be a possible correct vision of the future of coordinating agents to code with”
Comment from 15 Apr 2026 in “Gas Town: From Clown Show to v1.0” “Orchestration finally clicked with me when I tried out Gas Town six months ago”
Comment from 22 Aug 2026 in “Munder Difflin – Agent harness to run an office of your clones” Works in practice, produces lots of code
“Gas Town can pump out a huge amount of "working" code in short timeframes at obscene costs.”
Comment from 24 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale” “it implemented exactly what I asked for in a workable fashion with effectively one prompt”
Comment from 19 Jan 2026 in “Gas Town Decoded” “i think this is the real genius of gastown— not how it works, but that it does work”
Comment from 27 Mar 2026 in “Anatomy of the .claude/ Folder” - Comment from 6 Mar 2026 in “Labor Market Impacts of AI”
Interesting experiment worth exploring
“I think that's interesting, and I'm glad that someone built a (somewhat) working system to show what happens if you do that”
Comment from 27 Jan 2026 in “The age of Pump and Dump software” “Gastown, by Steve Yeggs is that, via tmux. It's rather opinionated and still in development, but it's worth a look”
Comment from 6 Jan 2026 in “Stop Doom Scrolling, Start Doom Coding: Build via the terminal from your phone” “Bravo to Steve for getting these thoughts on paper and the idea built even in such crude form”
Comment from 23 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale” - Comment from 19 Jan 2026 in “Gas Town Decoded”
Likely influences future agent systems
“gas town, or a successor built on a similar architecture, will not only be useful, but likely be considered 'state of the art'”
Comment from 19 Jan 2026 in “Gas Town Decoded” “Gastown preceded real support for multiagent orchestration in Claude and potentially nudged it along that path.”
Comment from 16 Apr 2026 in “Does Gas Town 'steal' usage from users' LLM credits to improve itself?” “Gas Town is a very instructive example of what the future of AI coding will look like”
Comment from 23 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale” - Comment from 23 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale”
What people warn about
Each mark is one comment
Expensive to run
“My main complaints with Gastown are that (1) it's expensive, partly because (2) it refuses to use anything but Claude models”
Comment from 7 Apr 2026 in “Google Open Sources Experimental Agent Orchestration Testbed Scion” “I do find the idea interesting, but heeded it's advice when it said it's hideously expensive and risky to use”
Comment from 14 Apr 2026 in “Gas Town: From Clown Show to v1.0” “Gas Town can pump out a huge amount of "working" code in short timeframes at obscene costs.”
Comment from 24 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale” - Comment from 27 Aug 2026 in “CEO fired developers to make room for AI. Developers create open source AI CEO”
Vibe-coded, unreviewed code
“Great tool, but the 100% vibecoding without looking at the code, for something that you are actually expecting others to use, is a bad idea.”
Comment from 23 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale” “Looking at something like the Beads and Gas Town repos, which are apparently fully vibe-coded”
Comment from 11 Jan 2026 in “AI Coding Assistants Are Getting Worse” “It's an open source 21st century master piece of ai psychosis, slop, and most especially redundant processes, orchestration and code”
Comment from 30 May 2026 in “The Last Technical Interview” Unnecessary complexity and redundancy
“No real software has been produced by the concept to date, except more garbage software that takes hundreds of thousands of lines where a few thousand would do.”
Comment from 14 Apr 2026 in “Gas Town: From Clown Show to v1.0” “people will still revel into inventing crazy architectures full of arcane terminology, independent of how much it's actually needed”
Comment from 31 Jan 2026 in “Coding Is When We're Least Productive” “I guess Gas Town's multiple layers of supervisory entities are meant to replace this manual tidying and fixing, but, well, really?”
Comment from 23 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale” - Comment from 14 Apr 2026 in “Gas Town: From Clown Show to v1.0”
Dangerous and largely untested
“If we must have GasTown/City/Metropolis then at least get an agent to examine and block potentially harmful commands”
Comment from 26 Apr 2026 in “An AI agent deleted our production database. The agent's confession is below” “Gas Town is an industrialized coding factory manned by superintelligent robot chimps, and when they feel like it, they can wreck your shit in an instant.”
Comment from 23 Jan 2026 in “Gas Town's Agent Patterns, Design Bottlenecks, and Vibecoding at Scale” “This tool is dangerous, largely untested, and yet may be of interest if you are already doing similar things in production.”
Comment from 6 Jan 2026 in “Welcome to Gas Town”
Other agent orchestrators people recommend
The four most recommended other agent orchestrators, 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 |
|---|---|---|---|---|---|---|
| Gas TownThis page | 95 | 212 | 31% | 10 | 1 mo3 owners | — |
| Conductor | 58 | 25 | 70% | 11 | — | Approved |
| Superset | 33 | 16 | 67% | 8 | 2 mo4 owners | Approved |
| Scion by Google | 28 | 32 | 47% | 8 | — | — |
| Agent Swarm | 16 | 3 | 84% | 6 | 8 mo3 owners | Approved |
What people said
40 of the 367 comments that mention it, newest first. Each links to the original comment.
“They're both expensive to operate, because the AI members of the organization spend a lot of time talking with each other.”
“the idea of having orchestrator threads to manage your work and keep tabs on it, having validators to validate the other work etc. were generally the right shape”
“Orchestration finally clicked with me when I tried out Gas Town six months ago”
“it seems to ignore the the more fundamental challenge of getting the human idea into software form”
“I have seen people with that kind of AI assisted productivity-by-volume, building eldritch abominations like Gas Town.”
“It's an open source 21st century master piece of ai psychosis, slop, and most especially redundant processes, orchestration and code”
“If we must have GasTown/City/Metropolis then at least get an agent to examine and block potentially harmful commands”
“Gastown preceded real support for multiagent orchestration in Claude and potentially nudged it along that path.”
“This is actually the most sustainable approach to maintaining open-source software that we've seen so far”
“it does seem to be a possible correct vision of the future of coordinating agents to code with”
“No real software has been produced by the concept to date, except more garbage software that takes hundreds of thousands of lines where a few thousand would do.”
“I do find the idea interesting, but heeded it's advice when it said it's hideously expensive and risky to use”
“this gave me the very clear feeling that most of the complexity of gas town is absolutely not needed and probably detrimental”
“I have otherwise never had an issue when I defined what the gate is, Claude usually honors it.”
“it generates a lot of code very quickly, that code is probably only ever supposed to be LLM maintained, not by people”
“My main complaints with Gastown are that (1) it's expensive, partly because (2) it refuses to use anything but Claude models”
“I'm skeptical that if I forked a vm running a busy Gas Town that it would be very light or fast in how it forks.”
“i think this is the real genius of gastown— not how it works, but that it does work”
“the volume of code and apparent utility (based purely on popularity) is undeniable”
“I think it's cool that folks like Steve with lots of money and time are using it to build stuff they find cool and interesting”
“Not reporting great success yet but also probably work that just wouldn't be done otherwise.”
“I would never seriously consider using for anything productive, but I love that he just put it out there”
“people will still revel into inventing crazy architectures full of arcane terminology, independent of how much it's actually needed”
“I think that's interesting, and I'm glad that someone built a (somewhat) working system to show what happens if you do that”
“it has a different type of worker that merges work and if needed can creatively rewrite completed work”
“Gas Town can pump out a huge amount of "working" code in short timeframes at obscene costs.”
“better enjoyed as more of Fear And Loathing-style ACID-fueled fevered dream than as a productivity tool”
“Bravo to Steve for getting these thoughts on paper and the idea built even in such crude form”
“I think ideas from it will probably partially inspire future, simpler systems.”
“Gas Town is an industrialized coding factory manned by superintelligent robot chimps, and when they feel like it, they can wreck your shit in an instant.”
“Gas Town is a very instructive example of what the future of AI coding will look like”
“Great tool, but the 100% vibecoding without looking at the code, for something that you are actually expecting others to use, is a bad idea.”
“I guess Gas Town's multiple layers of supervisory entities are meant to replace this manual tidying and fixing, but, well, really?”
“gas town, or a successor built on a similar architecture, will not only be useful, but likely be considered 'state of the art'”
“It is worth an install; it works very differently than an agent in a single loop.”
“it implemented exactly what I asked for in a workable fashion with effectively one prompt”
“The waste and lack of control is wild, but at the same time there's at least a nugget of fascinating, useful work in there.”
“Looking at something like the Beads and Gas Town repos, which are apparently fully vibe-coded”
“Gastown, by Steve Yeggs is that, via tmux. It's rather opinionated and still in development, but it's worth a look”
“This tool is dangerous, largely untested, and yet may be of interest if you are already doing similar things in production.”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.