Hi all,
I’ve just released DataGroceries 1.0: open-source Django apps that people and AI agents work on together. There are two so far, a kanban board and a wiki, and each installs on its own into a project you already have.
What they do
- The board: groups with their own columns, tickets with parents, blockers and relations, comments, @-mentions and watchers, a history of every change, and links to the GitHub commits and pull requests that name a ticket.
- The wiki: pages in one tree,
[[links]]between pages that follow a page when it moves, every version kept and comparable, comments, search, and a page in several languages.
The part I care about most: agents as first-class users. Every app has a REST API and tools on one MCP endpoint, and both run the same code as the pages. An agent such as Claude connects as the person who added it, through your own login. It reaches only what that person may use, follows the same rules, notifies the same people, and the history says “via Claude”. There are no agent accounts and no shared keys.
It fits the project you have. Your user model, your login and your database stay as they are. Who counts as your people, what they’re called and who may use what are hooks you can replace, with Django’s permissions as the default. Django 5.2 and 6.x, Python 3.11 and newer, MIT.
pip install django-dg-board django-dg-wiki django-dg-connect
- Try it in the demo, with Claude connected in a minute: https://demo.datagroceries.com
- Code and the integration guide: GitHub - datagroceries/django-dg: Open-source Django apps that people and AI agents work on together: a kanban board and a wiki, connected through one MCP endpoint. · GitHub
The board started as the internal tracker of Eddy Grid, who kindly agreed to its release. I run my own work on it, with Claude working the tickets alongside me.
I’d love to hear how it fits, or doesn’t, into your projects, and which app you’d want next.