Show HN: Vibe Kanban – Kanban board to manage your AI coding agents (github.com)

I used this last week and it's excellent - feels like the same increase in productivity increase from when I first used Cursor.

Are you thinking of doing a hosted version so I can have my team collab on it?

And I found I could open lots of PRs at once but they often need to be dependent on each other - and then I want to make a change to the first one. How are you thinking of better managing that flow?

How does this compare with Backlog.md? [1]

[1]: https://news.ycombinator.com/item?id=44483530

I tried to build something similar but in a peer-to-peer fashion and for humans + AI. It was supposed to be like a Kanban board that could scale to any team size and use Planning AI to ingest/match/monitor work realtime across teams and agents. I ran out of steam and couldn’t get funding but here is the prototype version:

https://postwork-alpha.vercel.app/

User: maryann.biaggioli@astarconsulting.com

Pass: Test1234!

I never got to a point where I actually integrated AI agents (weren’t as good at the time) but it’s cool to see it working in the real world!

I built something similar for my own workflow. Works okay. The hard part is as you scale, you end up with compounded false affirmatives. Model adds some fallback mechanism that makes it work, tests pass, etc. The nice part is you can ask models to review the code from others, call out fallbacks, hard coding, stuff like that. It does a good job at identifying buried bodies. But if you dig up a buried body, I'd manually confirm it was properly disposed of as the models usually hid the body in the first place because they needed some input they didn't have, got confused or ran into an issue.

This is a launch by a YC company that converts enterprise cobol code into java. Maybe it's my fault, but i tried every single coding agent with a variety of similar tools and whenever i try to parallelize, they clash while editing files simultaneously, i lose mental context of what's going on, they rewrite tests etc.

It's chaos. Thats fine if you are vibe coding an unimportant nextjs/vercel demo, but i'm really sceptical of all this stance that you should be proud of how abstracted you are from code. A kanban board to just shoot off as many tasks as possible and just quickly read over the PR's is crazy to me. If you want to appear a serious company that should be allowed to write enterprise code, imo this path is so risky. I see this in quite a few podcasts, tweets etc. People bragging how abstracted they are from their own product anymore. Again, maybe i am missing something, but all of this github copilot/just reviewing like 10 coding agents PR's just introduces so much noise and slop. Is it really what you want your image to be as a code company?

I remember the original Bloop search engine!

Some kind of UI or management system like this seems like it would be high level useful. Will have to give it a run.

Now we just let a CTO agent create the cards, review, and merge the PRs?

Very cool and interesting project. Ideas like this are a threat to traditionally-conceived project management platforms like Linear; that being said, Linear and others (Monday, ClickUp, etc.) are pushing aggressively into UX built for human/AI collaboration. I guess the question is how quickly they can execute and how many novel features are required to properly bring AI into the human project workspace

We need something like a kitchen brigade in software - one who writes the vibe code tickets (Chef de Vibe), one who reviews the vibe code (Sous-Vibe), one who oversees the agents and restarts them if they get hung up (Agent de Station). We could theoretically smash thousand tickets a day with this principle

Why does this need GitHub auth? This asks for unlimited private access to ones repo. This is a hard NO from me.

> Get 10x more out of...

So you're saying it goes up to 11x?

Why do you need to "manage" your coding agents like they are people? How long does it take them to do a task once prompted, in the background?

Don't you just prompt an immediately review the result?

Louis did a long form chat and vibe coded extra features in our discord meetup: https://www.youtube.com/watch?v=NCksand7Iwo for more info on this

The permissions this asks for feel kinda insane to me. Why does a kanban board need to see the code or my deploy keys among other things?

> human engineers now spend the majority of their time planning, reviewing, and orchestrating tasks

This feel like much too broad a statement to be true.

great. now i have to spend my morning using vibe kanban to make a tui for vibe kanban with textual. i'll submit a pr as soon as it's finished.

Would love to have a good selection of keyboard shortcuts! Power Vibe Kanban

[flagged]

That's an interesting idea and looks neat! I have my own developed agent running locally in containers, and currently use GitHub issues+pull requests for coordinating all the asynchronous work. Do you have any pointers on the approach I should take if I basically have something like a service already running for this, and I just want to hook up your UI to use it instead? Just some broad pointers on what would be required would be most helpful already!

I definitely don't feel like the models are reliable enough that I'd be more productive running them in parallel like this yet, but I can see a future where I want this.

Their reliability probably varies a lot depending on what you are using them for - so maybe I'm just using them in more difficult (for claude) domains.

Do you think you will keep it free or can you see a business model developing around it? If so, what do you think it would be? / How would you split paid tiers vs free users? Not a big deal to me...!! But I'm curious how one might commercialise these types of free/open source projects

Can you use this with the Claude Code Github Action?

Hmm, analytics appear to default to enabled: https://github.com/BloopAI/vibe-kanban/blob/609f9c4f9e989b59...

It is harvesting email addresses and github usernames: https://github.com/BloopAI/vibe-kanban/blob/609f9c4f9e989b59...

Then it seems to track every time you start/finish/merge/attempt a task, and every time you run a dev server. Including what executors you are using (I think this means "claude code" or the like), whether attempts succeeded or not and their exit codes, and various booleans like whether or not a project is an existing one, or whether or not you've set up scripts to run with it.

This really strikes me as something that should be, must legally be in many jurisdictions, opt in.

If you use gitlab, you can use the command line "glab" tool to have agents work from the built in kanban. They can open and close tasks, start MRs off of them etc. It's not as integrated as this tool, but works well with a mix of humans and robots.