People first · technology second
Shaper OS: the OS, then your tool
Shaper OS is the pillar. On it we shape the tool for your activity — clients, stock, shop, association — so you can run it day to day without becoming an engineer. It is not a chatbot, and not a frozen monthly rental: it is a system of your own, that keeps going when something breaks, and that we know how to raise again. The technique is there; it only shows itself to those who want to see the engine.
Two profiles, one product
Business owner or technical lead: read the card that fits you. No need to go through everything.
I want it to work day to day
You want to see your clients, your files and your activity — not administer servers. We install the pillar, adapt the tool to your trade, you run it day to day.
I want to check the tech
Where the code runs, who accesses what, which AI models, backups and recovery after failure. Detail below, in “under the hood”.
What it is — and what it is not
A pillar, then your trade
The same core serves a tradesperson, an SME, an association. What changes is the tool shaped on top: follow-ups, stock, shop, members, planning.
Not ChatGPT in a tab
An assistant wired to your context can act: file, follow up, prepare, alert. Screens stay optional. It is not a frozen product everyone rents the same way.
Not one cloud, not one model
We choose with you where data goes: a vendor, hybrid, or yours. A local model is possible when the case requires it. We do not promise magic equality with any model — we choose by use.
You stay the owner
The tool is yours. We can run it for you, or hand it over. Shaping is the method; Shaper OS is the pillar.
It keeps going — even when a piece fails
In an owner’s language, this is business continuity: the tool does not stop for an update, and we do not wipe your data each time it evolves. Backups sit at several levels — the machine, the useful files, the database, the code, and an off-site copy. With all five, you are covered; a missing level is a hole.
We know how to raise it from zero — and we prove it
This is disaster recovery: if the machine disappears, we rebuild the system and put your data back. It is not a slogan. We test it on an empty machine, then destroy the test — otherwise it is no longer a proof.
Restore time is not one number. Three distinct cases combine: images already cached; images to rebuild or pull from zero; then putting data back, which grows with your volume. The detail is under the hood. We refuse a magic SLA such as “under 120 seconds” for everyone.
Under the hood — for those who want to know or understand the engine
This section describes the technical side: where the code runs, who accesses what, which AI models, backups and recovery. An owner can skip it. A CIO or CTO should be able to read it without a meeting.
Runtime contract
Linux + containers (Podman). Laptop, VPS, a machine on your premises, or an isolated Linux delivered as a workstation. Not a magic Windows app. We can operate it for you if you do not want to touch the machine.
Not one vendor, not one agent
The pillar remains if we change model or agent (bridges to OpenCode, Cursor, Claude, and others). Frontier quality is not guaranteed identical on a local model: we choose by use, we do not bluff.
Case 1 — images already cached
Container images are already in our registry (or local cache). We redeploy the system without rebuilding everything: this is the shortest path for the technical layer. It still says nothing about how much data you have.
Case 2 — images rebuilt or pulled from zero
Cold registry, new machine, or images to rebuild: we must build or pull before starting. That time stacks onto the deploy — it is not the same clock as case 1.
Case 3 — putting data back
Once the system is up, we restore files, database and volumes. That delta grows with your volume. The three cases combine by scenario — we do not publish an “under 120 seconds” figure as a single SLA for everyone.
Five backup levels
Entire container (infra); persistent volumes as a compressed archive (turbinobash spirit, adapted to Podman, cache excluded); database dumps; tagged git (code, not your files); encrypted off-site copy (S3 / R2). All five together are enough.
A strong GDPR path, not a stamp
A local model (Ollama and equivalents) avoids sending prompts to a giant. GDPR is also retention, access, remaining processors (voice, tunnel, mail) and the pillar’s audit log. A lawyer still belongs on sensitive cases.
Public kit, operator proof
The install kit is public. The CRM demo runs and resets every night. You can read the law, not only the brochure.
Two public proofs
The Shaper OS V1.6 GitHub kit: clone, install, a system law you can audit. And the CRM demo: you go in, you click, data resets every night.
Frequently asked
- Do I need to understand IT to use it?
- No. You talk about your trade; we shape the tool. The technical section above is for those who want to inspect — not an entry requirement.
- Is this like Claude Cowork or a WhatsApp agent?
- No. Those tools are agents. Shaper OS is the activity system: your business tool on top, continuity, recovery, isolation. We can deliver an isolated Linux workstation; it is not a sandbox folder at a vendor.
- Where does it actually run?
- On a Linux machine — yours, ours, or an isolated Linux we deliver. Cloud is not forbidden: it is a cursor, not a religion. The sovereignty detail is on the AI page.
Curious whether this is for you?
Owner or CIO: we talk about your activity first, not the tech. Thirty minutes to see if the subject resonates — no strings, and what comes next is decided together.