[HN Gopher] Show HN: Vibe Kanban - Kanban board to manage your A...
___________________________________________________________________
Show HN: Vibe Kanban - Kanban board to manage your AI coding agents
Hey HN! I'm Louis, one of the creators of Vibe Kanban. We started
working on this a few weeks ago. Personally, I was feeling pretty
useless working synchronously with coding agents. The 2-5 minutes
that they take to complete their work often led me to distraction
and doomscrolling. But there's plenty of productive work that we
(human engineers) could be doing in that time, especially if we run
coding agents in the background and parallelise them. Vibe Kanban
lets you effortlessly spin up multiple coding agents. While some
agents handle tasks in the background, you can focus on planning
future work or reviewing completed tasks. After a few weeks of
internal dog fooding and sharing it with friends, we've now open-
sourced Vibe Kanban, and it's stable enough for day-to-day use.
I'd love to hear your feedback, feel free to open an issue on the
github and we'll respond ASAP.
Author : louiskw
Score : 140 points
Date : 2025-07-11 15:08 UTC (7 hours ago)
(HTM) web link (github.com)
(TXT) w3m dump (github.com)
| lharries wrote:
| 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?
| louiskw wrote:
| Yeah I think giving the option to move execution to the cloud
| makes a lot of sense, I already find my macbook slowing down
| after 4 concurrent runs, mainly rustc.
|
| Also now we're pushing many more PRs think we defo need better
| ways to stack and review work. Will look into this asap
| hddbbdbfnfdk wrote:
| Very productive increase sirs! Whole team well promoted.
| slig wrote:
| How does this compare with Backlog.md? [1]
|
| [1]: https://news.ycombinator.com/item?id=44483530
| ggordonhall wrote:
| Hi, co-author of this project here!
|
| In Vibe Kanban you can directly interact with Claude Code from
| within the Kanban board. E.g. you can write out a ticket, hit a
| button to run it locally with Claude Code/Gemini etc., watch
| its responses, and then review any diffs that it generated.
| slig wrote:
| Thank you! I was taking a look at the docs, and I'm going to
| play with it later today. Thanks for sharing and congrats on
| shipping!
| randysalami wrote:
| 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!
| swalsh wrote:
| 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.
| deepdarkforest wrote:
| 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?
| unshavedyak wrote:
| > 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.
|
| Fwiw Claude suggests using different git workspaces for your
| agents. This would entirely solve the clashing, though they may
| still conflict and need normal git conflict resolves of course.
|
| Theoretically that would work fine, as it would be just like
| two people working on different branches/repos/etc.
|
| I've not tried that though. AI generates way too much code for
| me to review as it is, several subtasks working concurrently
| would be overwhelming for me.
| louiskw wrote:
| This is a bet on a future where code is increasingly written by
| AI and we as human engineers need the best tools to review that
| work, catch issues and uphold quality.
| deepdarkforest wrote:
| I don't disagree, but the current sentiment i was referring
| to seems to be "maximize AI code generation with tools
| helping you to do that" rather than "prioritize code quality
| over AI leverage, even if it means limiting AI use somewhat."
| codingdave wrote:
| It is not just chaos, it is an unwanted product. Don't
| misunderstand - people would love this product if it works. But
| AI cannot do this yet. Products like this are built on an
| assumption that AI has matured enough to actually succeed at
| all tasks. But that simply isn't true. Vibe coding is still
| slop.
|
| AI needs to do every single step of this type of flow to an
| acceptable quality level, with high standards on that
| definition of "acceptable", and _then_ you could bring all the
| workflow together. But doing the workflow first and assuming
| quality will catch up later is just asking for a pile of
| rejections when you try to sell it.
|
| I'm not just making this up, either... I've seen and talked to
| numerous people over the last couple years who all came up with
| similar ideas. Some even did have workable prototypes running.
| And they had sales from the mom/friends/family connections, but
| when they tried to get "real" sales, they hit walls.
| skeptrune wrote:
| 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.
| louiskw wrote:
| Nostalgia!
|
| Let me know if any issues, we're turning feedback around pretty
| quick
| shibeprime wrote:
| Now we just let a CTO agent create the cards, review, and merge
| the PRs?
| ggordonhall wrote:
| We have this! Vibe Kanban includes its own MCP server that you
| can use to create tickets within the Kanban board.
|
| Click on the MCP Servers tab, then hit "Add Vibe Kanban MCP".
| Then create and start a "planning" ticket like "Plan a
| migration from AWS to Azure and create detailed tickets for
| each step along the way". Sit back and watch the cards roll in!
|
| Will do more to document this better soon :)
| _jayhack_ wrote:
| 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
| louiskw wrote:
| Cheers! Smaller teams, more infrastructure, more testing, tasks
| requiring review in minutes not days - the features are just
| totally different for the new world than what legacy PM tools
| are optimised for, and who they have to continue to serve.
| oc1 wrote:
| 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
| ggordonhall wrote:
| Completely agree!
|
| You can actually use a coding agent to create tickets from
| within Vibe Kanban. Add the Vibe Kanban MCP server (from MCP
| settings) and ask the agent to plan a task and write tickets.
| csomar wrote:
| Why does this need GitHub auth? This asks for unlimited private
| access to ones repo. This is a hard NO from me.
| louiskw wrote:
| It can open GitHub PRs from the interface and there's tons more
| info we want to pull in like the result of CI checks
| bpshaver wrote:
| > Get 10x more out of...
|
| So you're saying it goes up to 11x?
| louiskw wrote:
| Great ChatGPT question XD
| skeeter2020 wrote:
| AI was already giving you 100x so we're up to 1000x
| remram wrote:
| 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?
| scotty79 wrote:
| I don't think smart ones are that fast. They can work for hours
| if you have the budget.
| remram wrote:
| Interesting, I didn't know that.
| louiskw wrote:
| For me, the average task takes a coding agent 2-5 minutes to
| complete. A slightly annoying amount of time as I'm prone to
| getting distracted while I wait.
|
| This gives me something to do in that time.
|
| My guess is time to complete a task will oscillate - going up
| as we give agents more complex tasks to work on, and going down
| with LLM performance improvement.
| gpm wrote:
| Currently - claude code is pretty slow. I've definitely had it
| take >15 minutes on the (faster than opus!) sonnet model just
| thinking and writing code without feedback or running long
| lasting tools. I expect this to change given that companies
| like cerebras exist and seem to know how to generate tokens in
| much less time, but the current state of the art is what it is.
|
| Always - if you're going to pipe the result of some slow
| process back to them (like building a giant C++ project that
| takes minutes/hours, or running a huge set of tests...)... it's
| going to be slow.
| swyx wrote:
| 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
| louiskw wrote:
| "11 days ago" this feels like a year ago :O
| iimblack wrote:
| 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?
| jeltz wrote:
| I would assume because it was vibe coded.
| gpm wrote:
| More generously I'd assume because
|
| - It's an early prototype so they haven't dealt with fine
| grained permissions
|
| - They really do want to do things like access private repos
| with it themselves
|
| - They really do want the ability to do things like checkout
| code, create PRs, etc... and that involves a lot of
| permission.
| skeeter2020 wrote:
| every one of your "more generous" assumptions is the
| opposite of what should be their process. It's the
| equivalent of "vacuum up as much data as possible and then
| decide what to do with it". Not acceptable.
| gpm wrote:
| It's "vacuuming" that data in the sense of giving API
| access to a tool that runs on your local computer, that
| seems acceptable enough for me in the early stages of
| developing a tool.
|
| The other privacy complains I have regarding them
| harvesting usernames and email addresses... not so much.
| TeMPOraL wrote:
| Because _it 's not_ "a kanban board"? It's a coding agent
| orchestrator that's made in the shape of a Kanban board.
|
| You might be right that this app asks for excessively broad
| privileges, but your case would be much stronger if it wasn't
| backed by an absurdly disingenuous argument.
| barbazoo wrote:
| > 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.
| bwfan123 wrote:
| > AI coding agents are increasingly writing the world's code
| and human engineers now spend the majority of their time
| planning, reviewing, and orchestrating tasks.
|
| This tactic is called "assuming the sale". ie, Make a statement
| as-if it is already true, and put the burden on the reader to
| negate it. Majority of us are too scared of what others think,
| and go-along by default. It is related to the FOMO tactic in
| that it could be used in conjunction with it to make it a
| double-whammy. for example, the statement above could have
| ended with: "and everyone is now using agents to increase their
| productivity, and if you arent using it, you are left behind"
|
| Glad you stood up to challenge it.
| skeeter2020 wrote:
| I'll add - often not adding the last part is even MORE
| powerful: "and everyone is now using agents to increase their
| productivity..."
| ljm wrote:
| I wouldn't say it's the majority of my time but the most
| utility I've got out of AI is using MCP to deal with the boring
| shit: update my jira tickets to in progress/in review, read
| feedback on a PR and address the trivial shit, check the CI
| pipeline and make it pass if it failed, and write commits in a
| consistent, descriptive way.
|
| It's a lot more hands on when you try to write code with it,
| which I still try out, but it's only because I know exactly
| what the solution is and I'm just walking the agent towards it
| and improving how I write my prompts. It's slower than doing it
| myself in many cases.
| rvz wrote:
| I read that too and these are the kind of statements which
| really tells you what happens when a profession embraces
| mediocrity and accepts something as crass as "Vibe-coding"
| which is somehow going to change "software engineering" even
| when adding so-called "AI agents" - which makes it worse.
|
| All this cargo-culting is done without realizing that more code
| means more security issues, technical debt, more time for
| humans to review the mess and *especially* more testing.
|
| Once again, Vibe-coding is _not_ software engineering.
| skeeter2020 wrote:
| and I came into the industry when software was not
| engineering. Still think this is mostly true (you can call
| yourself an engineer when you insure your product)
| lazarus01 wrote:
| > 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.
|
| This is just what they wish to be true.
| dhorthy wrote:
| i feel so strongly that this will rapidly become true over the
| next 6 months. if you don't believe me check out Sean Grove's
| talk from mid June -
| https://www.youtube.com/watch?v=8rABwKRsec4
| lbrito wrote:
| I wonder how demographics (specifically age) tie into this. I'm
| well into my 30s and I found that statement absurd, but perhaps
| it is basically universally true among recent grads.
| bigfishrunning wrote:
| Maybe it is -- the next few years are going to get really
| rough for them; they'll develop no skills outside of AI.
| amirhirsch wrote:
| 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.
| Jolter wrote:
| It says in the linked readme to not open pull requests without
| discussing it first.
| peadarohaodha wrote:
| Would love to have a good selection of keyboard shortcuts! Power
| Vibe Kanban
| louiskw wrote:
| Good point - there are a few random ones already but I will ask
| a coding agent to implement this comprehensively
| lowdownbutter wrote:
| [flagged]
| diggan wrote:
| 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!
| gpm wrote:
| 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.
| louiskw wrote:
| Yes I generally cherry pick the easier 50% of my backlog and
| work on those with Vibe Kanban, and the other 50% is still
| manual or happens with coding agent but with a human-in-the-
| loop.
|
| This is a bet that coding agents will continue to get better,
| and this feels like the right time to try and figure out the
| interface.
| FailMore wrote:
| 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
| louiskw wrote:
| I could see there being a long term free offering that doesn't
| cost us compute or tokens, and probably some other offerings
| that actually do use resources and would make sense to build a
| business around.
|
| But that's not a today problem, we just want to absorb feedback
| and iterate until we build the ultimate tool for working with
| these coding agents.
| MH15 wrote:
| Can you use this with the Claude Code Github Action?
| gpm wrote:
| 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.
| bn-l wrote:
| Thanks, really appreciate the heads up. I put devs who do this
| on a personal black list for life.
|
| I think also that this would be better as an mcp tool /
| resource. Let the model operate and query it as needed.
| willsmith72 wrote:
| It's the email/username harvesting that you mean right? Or do
| people also have something against anonymised product
| analytics?
| gpm wrote:
| I have something against opt-out analytics over TCP/IP or
| UDP/IP period, because they aren't anonymized, they include
| an IP address by virtue of the protocol.
|
| But I definitely only posted that original complaint of the
| email/username (not the person you responded to initially).
| jjangkke wrote:
| analytics stuff is fine but the email harvesting/github
| username appears to be illegal especially if its done without
| notifying the user?
|
| great catch, many open source projects appear to be just an
| elaborate lead gen tool these days.
| swyx wrote:
| could you point me to what jurisdictions require analytics opt
| in esp for open source devtools? thats not actually something
| ive seen as a legal requirement, more a community preference.
|
| eg ok we all know about EU website cookie banners, but i am
| more ignorant about devtools/clis sending back telemetry. any
| actual laws cited here would update me significatnly
| gpm wrote:
| I mean, you've labelled one big one already with the GDPR
| covering a significant fraction of the world - and unlike
| your average analytics "username and email address" sounds
| unquestionably identifying/personal information.
|
| Where I live I think this would violate PIPEDA, the Canadian
| privacy law that covers all business that do business in any
| Canadian province/territory other than BC/Alberta/Quebec
| (which all have similar laws).
|
| There's generally no exception in these for "open source
| devtools" - laws are typically still laws even if release
| something for free. The Canadian version (though I don't
| think the GDPR does) has an exception for entirely non-
| commercial _organizations_ , but Bloop AI appears to be a
| commercial organization so it wouldn't apply. It also
| contains an exception for business contact information - but
| as I understand it that is not interpreted broadly enough to
| cover random developers email addresses just because they
| happen to be used for a potentially personal github account.
|
| Disclaimer: Not a lawyer. You should probably consult a
| lawyer in the relevant jurisdiction (i.e. all of them) if it
| actually matters to you.
| generalizations wrote:
| > GDPR covering a significant fraction of the world
|
| > privacy law that covers all business that do business in
| any Canadian province
|
| A random group of people uploaded free software source code
| and said 'hey world, try this out'. I wish the GDPR and the
| PIPEDA the best of luck in keeping people from doing that.
| (Not to actually defend the telemetry, tbh that's kinda
| sleezy imo.)
| gpm wrote:
| I mean, those are merely the two countries privacy laws
| I'm most familiar with. The general principal of "no you
| can't just steal peoples personal information" is not
| something unique to the ~550 million people the laws I
| cited cover.
|
| And the laws don't prevent you from uploading "random"
| software and saying "try this". They prevent you from
| uploading _spyware_ and saying "try this". Edit: Nor
| does the Canadian one cover any random group of people,
| it covers commercial entities, which Bloop AI appears to
| be.
| janoelze wrote:
| fork, task claude to remove all github dependence, build.
| gpm wrote:
| I did this locally to try it out :) Also stubbed out the
| telemetry and added jj support. "Personalizing" software like
| this is definitely one of LLMs superpowers.
|
| I'm not particularly inclined to publish it because I don't
| want to associate myself with a project harvesting emails
| like this.
| janoelze wrote:
| yes, i was just doing/thinking the same, it was an
| interesting experience to sculpt a somewhat complex
| codebase to my needs in minutes.
| hsbauauvhabzb wrote:
| Use a telemetry backed tool to remove telemetry from another
| telemetry backed tool?
| janoelze wrote:
| it came to mind first, you're free to use whatever flavour
| of LLM floats your boat vibes your code.
| hsbauauvhabzb wrote:
| That doesn't change the naivety of the response.
| TeMPOraL wrote:
| There's telemetry you consent to, and telemetry you don't.
| Just because I'm fine with a tool like Claude Code
| collecting some telemetry, doesn't mean I'm fine with a
| different party collecting telemetry - and the two products
| being used together doesn't change it. It's not naive, it's
| simply my right.
| louiskw wrote:
| That's fair feedback, I have a PR with a very clear opt-in here
| https://github.com/BloopAI/vibe-kanban/pull/146
|
| I will leave this open for comments for the next hour and then
| merge.
| TeMPOraL wrote:
| Nice, I vote for merging it :).
|
| It really doesn't hurt to be honest about this and ask up-
| front. This is clear enough and benign enough that I'd
| actually be happy to opt-in.
| louiskw wrote:
| Merged and building, thanks for bearing with us
| gpm wrote:
| I concur :)
| uxamanda wrote:
| 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.
| louiskw wrote:
| Interesting, hadn't heard of that. Would better GitLab support
| be useful in Vibe Kanban?
| jjangkke wrote:
| I'm not sure if Kanban is the right UI for what this is supposed
| to be for, just a gut feeling. Curious what other UI is more
| appropriate for this.
| louiskw wrote:
| Kanban seems like a good starting place, but I broadly agree
| that the interface for human<>agent collaboration will need to
| be different from the default interface we have today with
| legacy PM tools.
|
| Things move across the board so quickly when AI is doing the
| work that ~50% of the columns seem pretty redundant.
| drbojingle wrote:
| I'm building one of these too :D
| skeeter2020 wrote:
| If I multiply my 100x productivity gains from using AI with your
| 10x increase what am I supposed to do with all that free time?
| ffsm8 wrote:
| Maybe Tony can inspire you?
|
| ( * _ ) --
|
| https://youtube.com/shorts/YBAcvRV7VSM?si=jp2hZvFIVo-vSdu6
| jackbridger wrote:
| This is interesting. I'm using Claude Code a fair bit and have
| found writing specs to be more effective than promoting and this
| feels closer to that. I can see the appeal of this for simple
| tasks now and maybe increasingly bigger tasks as models get
| better.
| louiskw wrote:
| Very much a bet that things are going to get much much better
| very quickly
| sqs wrote:
| This is really cool. I used Vibe Kanban with Amp to update some
| of our docs and UI components, and it was great.
| louiskw wrote:
| I would say conservatively that 80% of Vibe Kanban has been
| built by Amp
___________________________________________________________________
(page generated 2025-07-11 23:01 UTC)