Asking community and maintainers about posting my tool for everyone to flathub

1

Hello, I’m a developer and I’d like to ask before submitting, not after.

I’ve read the Generative AI policy, including the line about exceptions for
mature, well-maintained projects. My project is developed with AI assistance
and I state that openly in the README and in AGENTS.md, so I want to ask
directly whether it has any chance under that exception, and what bar you
would want to see.

The app is Unisic (GitHub - unisic/unisic: Screenshot & screen-recorder for Linux. · GitHub), a screenshot and screen
recording tool for Linux. GPL-3.0, no telemetry, no account, Wayland
first and working on X11 too. It uses only the supported capture paths:
xdg-desktop-portal, KWin ScreenShot2 on Plasma, PipeWire for video.

How it is built, honestly: I use AI agents to write code, and I review every
change line by line before it lands. I decide what gets built and what gets
rejected, I test on real desktops (KDE, GNOME, wlroots, COSMIC), and I fix the
bugs that come out of that. It is not a wrapper or a one-prompt app.

Where the project is today: first commit 2026-07-07, 341 commits, releases
From 0.1 (on git from 0.3) to 0.7.5, packages for Debian/Ubuntu, Fedora, openSUSE, and Arch
plus an AppImage, 10 unit tests, and an in-app smoke test, 5 interfaces
languages, and one accepted translation PR from an outside contributor. I know
Three weeks is not a long track record, which is exactly why I’m asking rather
than assuming.

Two things I want to flag. The Spanish and Italian translations
were machine-generated (English and Polish are written by me, and French came from a human contributor), and I read that
This is not accepted, so I plan to replace them. And the Flatpak manifest is
It is ready and passes flatpak-builder-lint, but I have not opened a submission PR.
because I would rather hear a no here than violate the policy there.

So my question is: is an exception possible for a project like this, and if so,
What would you need to see before it would be worth submitting?

Thanks for your time.

2

Only three weeks of development would also run into issues with another policy:

To be frank: We see too many submissions nowadays which were completely cobbled together using AI which are not of great quality. And most often they are immediately abandoned as well:

We are experiencing a wave of slop made by “developers” which have no idea or interest in software quality or maintenance. But who all feel entitled that what they let the chatbot statistically calculate has to be on Flathub.

Which is why we have these rules in place. So, currently I would deny your submission.

Once you have more development history for your application and a trustworthy plan of its continued maintainance, we can revisit this.

3

I use AI agents to write code, and I review every change line by line before it lands.

Every AI project, the devs swear they review very carefully. And every AI project i end up looking at the same weird nonsensical sludge that looks even less reviewed than random drafts on a napkin.

I am aware the policy mentions that exceptions can occur, but that is an escape hatch. Everyone think they are the exception, and the previous version let people to believe they could argue much more than now.

My bar is that if the project is still using AI or has an extensive track record of using it, then it is a liability to have it in, and i would immediately reject

And while i try to stay focused on the technical aspect, i cannot be fully objective, and on a personal level view these tools as deeply abhorrent morally. So theres also from the get go not going to be much clemency…