30 days to launch: inside Archibou

By Etienne, founder of Archibou, September 1st, 2026

We launch on October 1st. It's September 1st.

30 days left. It's the right time to tell you where we really stand, what's ready, what's not yet, and what we'll do with this last month. No sugar-coating.

If you've followed Archibou since July, you know we're building a sovereign alternative for your photos: your memories stored in France, on French infrastructure, with software you can understand and leave whenever you want. This post is not an ad. It's a status update.

What's ready and frozen

Let's start with the reassuring part: the core is there, and we won't touch it anymore.

Infrastructure. Deployment scripts for Clever Cloud are ready, functional and tested. Clever Cloud is a French hoster running its own infrastructure physically in France, not a region of a US giant. The real production deployment is scheduled for next week. We spent time automating so we can redeploy cleanly, not tinker by hand.

The cloud. The backend that stores your photos is a slightly hardened Nextcloud: we removed a few unnecessary apps and features, and reworked the default settings to improve security from day one. Features are frozen. There may be visual tweaks left, but the substance won't move.

Android app. Also feature-frozen. I spent the whole month of July on it, built with Flutter precisely to make what comes next easier: the iOS app, which may not be ready at launch depending on how September and beta testing go, but will follow without starting from scratch. Today I'm hunting the last bugs, personally, with my wife as a ruthless tester. It's concrete: install, back up, find your photos, see memories from previous years, and on a map.

Marketing site. The site you see today (Django + Postgres under the hood) is broadly OK. Pricing is already public, no surprises to expect.

Checkout and customer dashboard. Coded, validated on test infrastructure. They work. What's missing is the real-world test: on the actual prod infra, with real people outside the company.

What's still to do, honestly

And that's where the 30 days matter.

1. Deploy prod for real. Scripts are ready, but prod is not yet live. It's planned for next week. Until we've run on it, we don't claim victory.

2. Test with real people. The checkout + dashboard must be tested by future beta testers on prod. Not by us, in a corner.

3. Polish the legal wording, not the substance. Terms, legal notices and privacy policy are written in substance. What's left is a final review of the wording. Important nuance: we're not chasing a checkbox, we want it to be readable.

4. Complete FAQ and tutorials. Well started, but incomplete. And that's normal: a useful FAQ is written with real questions from early users, not in our heads.

No hosting cost surprise to announce, there is none at this stage. The cost was time: a full month of July on the app.

Three sentences that kept us going in August

We talked about Archibou around us this summer. Three pieces of feedback keep coming back, and they explain better than we do why we're building this:

"Really interesting. If I'd had this when my phone was stolen, I wouldn't have lost my memories!"

"Nice feature to see photo memories from previous years. Like Facebook, but in complete trust, no tracking, no Big Tech."

"The map to see where the photos were taken, that's really cool!"

No jargon. Simple uses: not losing your memories and finding them again. That's exactly the scope we're freezing.

The next 30 days, concretely

• Week of September 8th: production infrastructure goes live.
• Mid-September: opening to beta testers, full journey, from signup to complete backup of photos and documents. And their restore.
• End of September: fixes from beta feedback, final FAQ/tutorials and legal wording.
• October 1st: public launch.

If beta testing reveals a blocker, we'll delay iOS rather than rush Android. The opposite would be dishonest.

Why these technical choices?

In short, because we want you to be able to verify.

• Django + Postgres for the site: solid, open source, no black box.
• Hardened Nextcloud for storage: same logic, we start from open software we tighten instead of a closed service you have to take on trust.
• Clever Cloud in France for hosting: French company, infra in France, French jurisdiction. Your photos don't take a detour to another continent for legal reasons.

And Flutter for the app: one codebase for Android today and iOS tomorrow, without doubling technical debt.

What we won't do before October 1st

We won't add features. It's tempting, at 30 days out, to slip in "just one small option." We're forbidding it. The list is frozen for a reason: every addition pushes testing, hence reliability. If an idea comes now, it goes to the "after launch" list, not to prod.

We also won't promise the iOS app for day one. Better one app that holds than two half-ready. Flutter precisely lets us not have to choose for long: the codebase is shared, work isn't lost.

Follow what's next

We post regularly here. Opinions, explainers and company progress.

If you want to follow all this without checking the blog every week, the easiest is to follow us where we already post: on LinkedIn and Facebook for example. And of course here on the blog.

Thanks to everyone testing, proofreading and asking questions since July. The most useful 30 days start now.

Etienne