A few weeks ago, we replaced Jira with a folder of markdown files. We've coined it faru and we've been using it every day since. It stuck. So we put it on GitHub.
Why we built it
It's not that we built a tool to give our AI agents a backlog. It's about AI agents making their work visible to us.
Our agents write markdown files into a backlog/ folder in our docs repository. A bit like people would use a Google Drive, except it's git. Our IDEs, agents and tools already have that folder open. Git already syncs it.
We didn't want to integrate with another platform. We wanted the board to just be part of the flow we already had. Open the folder, see the cards. Save a file, the board updates. That's it.
At fluado, that's also what we try to do for our customers. Don't bolt a new platform onto someone's workday. Fit into what's already there.
We put it on GitHub because it might be useful to someone. Or at least inspire them to build their own version.
What it actually is
One server.js file. Zero npm dependencies. You point it at a directory, it reads the markdown, renders 3 columns.
npx github:fluado/faru
Cards are folders. Each folder has markdown files with YAML frontmatter. Drag a card to move it. Click a title to rename it. Every change gets committed and pushed. The board polls the remote every 5 seconds. Two people looking at the same repo see the same board.
Click a card and you get a detail view with metadata, milestone progress, comments, and a list of all the files in that folder. Click any file and it opens in your editor. The board doesn't try to be an editor. Your IDE is the editor.
3 columns
faru has 3 columns. You can't add more.
I know. It's on purpose.
"In Review" is WIP. "Blocked" is a problem to fix, not a lane to park in. "Ready for QA" is a queue, and queues are where tickets go to die.
A card is either not started, being worked on, or done. If you need more states than that, read up about lean methodologies.
What we're not doing
We're not building a SaaS. We're not going to add Gantt charts or time tracking.
We also won't maintain a community around this. We don't have the bandwidth to review PRs or triage issues.
If you open an issue with a good idea, we might pick it up when it aligns with what we need. No promises.
Fork it, break it, make it yours.
Who this might be for
If you work with AI coding agents and you want them to manage work through files the way they manage code, faru does that.
If you're a small team, maybe just 2 people and some agents, and you want visibility without a subscription, this might work (it does for us). It runs locally. Your data stays on your machine and in your git repo.
If you're a team of 50, this probably isn't for you.
We built it for us. We shared it because why not.
Want your own AI agents? Talk to us. We build AI agents that actually work hand-in-hand with medium-sized businesses, without the bloat.
If you want to follow what we're building at fluado, find us on LinkedIn or Bluesky.
Top comments (0)