AI should serve people.
We are not here to sell you the future or to scare you about it. We are here to use it on purpose, on your business, where it earns its place. This page is what that commits us to.
-
You stay in charge.
Every automation we build stops and hands over to a person at the point where being wrong would cost you something. That line is drawn on purpose, and you get to move it.
-
No hype, no doom.
Most of what you have been told about AI is either a sales pitch or a panic. We test things, tell you what happened, and skip the theater in both directions.
-
For everyone, not the few.
This does not require a technical team or a big budget any more. If you can describe the work, it can probably be automated — and if it cannot, we will say so.
-
Own what you make.
It runs on your accounts, with your data, and you get the keys. There is no platform of ours you have to keep paying to keep working.
What that rules out
Principles are cheap until they cost you a sale. These have.
-
We do not sell you a seat in our software.
Everything is built on your accounts and handed over. Our incentive is to finish, not to keep you subscribed.
-
We do not take a cut of the tools we recommend.
No referral fees, no reseller margin, no partner tier. If we point you at something, it is because it is the right thing.
-
We do not automate a broken process.
Making a bad workflow run faster and unattended is how a small problem becomes an expensive one. Sometimes the answer is to fix the process and stop.
-
We do not pretend a bot is a person.
Your customers get told what they are talking to, and it hands over to a human the moment it should. That is not a limitation, it is the design.
Good use
-
Decoding lost text
Scrolls carbonized by Vesuvius, unreadable for centuries, recovered by models trained to see through the ash.
-
Protein folding
AlphaFold solved a 50-year problem and released nearly every known structure, for free.
-
Earlier detection
Models trained on medical images catch tumors at stages when treatment still works.
Bad use
-
Slop at scale
Unreviewed generated text and images flood search, schoolwork, and public debate until nobody can tell who meant it.
-
Surveillance dressed as help
Systems that watch workers, students, or citizens under the banner of efficiency, with no way to contest the score.
-
Stolen work, laundered
Models trained on other people's writing, art, and code, then sold back as a product with the source erased.
Consequences
Every shortcut has a residue. A model that writes your email also trains the next person to stop thinking before they send. A detector that flags a student as a cheater can wreck a year on a false positive. A hiring screen that ranks résumés will bake yesterday's bias into tomorrow's staff.
The cost is not only jobs. It is trust: in text, in photos, in institutions, in each other. Use that is sloppy at personal scale becomes political at civic scale.
Social impact
AI is already in medicine, science, schools, and the public square. Who it serves is still being decided, mostly by people who will not live with the downside.
tuaregs exists to put capable tools in more hands, with the adult warning attached: you stay responsible for what you ship. The work is to widen access without widening harm.
Which is why the library is free
Everything we know about using this well is published, in order, at no cost and behind no email form. If you read it and decide to do the work yourself, that is a good outcome. It is also the cheapest way for you to find out how we think before you pay us anything.