Privacy
You are being asked to put work you have not published yet onto someone else's computer. That deserves a straight answer, not a wall of clauses.
Last updated 17 August 2026. Gapnostic is a product of Kairos Solution.
The short version
- Your research is yours. We do not sell it, mine it, publish it, or train anything on it.
- No other researcher can see your workspace. That is enforced in the database, not by us remembering to be careful.
- We do not record what you ask the AI or what it answers. Only that a request happened, and what it cost.
- There are no adverts, no tracking pixels, no analytics, and no cookies beyond the one that keeps you signed in.
- Ask us to delete your account and we delete it, along with everything in it.
What we hold
If you joined the waitlist
The form asks for your name, email address, role, institution, field and country. It goes into a private spreadsheet we use to decide who to invite and in what order. It is not a mailing list we rent, sell or share, and the only messages you will get from it are about getting access.
If you have a workspace
| What | Why it exists |
|---|---|
| Your email address and, if you gave one, your name | To know whose workspace is whose, and to reach you about your account |
| Your password, stored only as an irreversible hash | To let you sign in. We cannot read it, and neither can anyone who steals the database |
| A sign-in token, also stored hashed, expiring after 30 days | So you are not asked to sign in on every page |
| Your research: projects, papers, abstracts, uploaded PDFs, matrices, notes, reports | It is the thing the product is for |
| Credits granted and spent, and the date you were last active | So the AI budget is fair and you can see where it went |
What we deliberately do not record
Gapnostic logs every AI request so you can see what you spent. That log holds the name of the task, the model used, how many tokens it took, and the estimated cost.
It does not hold what you asked or what came back. We could have stored those. It would have made debugging easier. We decided that a record of every question a researcher asked about their unpublished idea is not a thing worth having.
Where it lives
On a server in Frankfurt, Germany, rented from Render. The database and any PDFs you upload sit on a disk attached to that server. Nothing is copied to another machine for analysis.
The waitlist spreadsheet is held in Google Workspace. The gapnostic.com site itself is served by Netlify.
Who else sees any of it
Making the product work means a few things leave our server. Here is all of it.
| Who | What reaches them |
|---|---|
| Anthropic the AI |
The text being analysed: your topic, the abstracts and papers you asked about. Anthropic does not use API data to train their models. We pay for this, so you are not the product. |
| OpenAlex the paper database |
Your search terms, so it can find papers. The request identifies Gapnostic, never you. OpenAlex is told what was searched for, never who searched. |
| Publishers and doi.org | Nothing about you. When you ask for an open-access PDF, we fetch it from wherever it is published, the same as clicking the link yourself. |
| Render | They host the server, so technically the data sits on their hardware. Standard for any hosted service. |
| Google Fonts | The typefaces on this website load from Google, which means Google's servers see your IP address when you visit. We intend to serve the fonts ourselves and remove this. |
That is the complete list. There is no advertising network, no analytics provider, no data broker, and no "partner" arrangement of any kind.
The honest part about administrators
Most privacy policies imply that nobody at the company could look at your data. That is almost never true, and it is not true here either.
The app gives no one a way to open another person's workspace. Even an administrator asking for a project by its exact identifier is told it does not exist. But whoever runs the server can reach the database directly, because that is what running a server means.
So the real promise is not that it is impossible. It is that we do not, we have no reason to, and if we ever needed to touch your data to fix something broken, we would ask you first.
If that is not good enough for the work you do, and for some research it should not be, Gapnostic also runs entirely on your own computer, where none of this applies. Ask us.
How long we keep it
Your workspace stays until you ask us to remove it. We are not in the business of quietly deleting a researcher's work because they were away for a term.
Sign-in tokens expire after 30 days. Waitlist entries are kept until the beta is over, then deleted.
Getting your data, or getting rid of it
Email ruhu@gapnostic.com and we will act within 30 days, usually the same week. You can ask us to:
- Show you everything we hold about you.
- Correct anything wrong.
- Delete your account and everything in it. This is not a soft delete or a hidden flag. The rows go, the files go, and it cannot be undone, so please be sure.
- Take it elsewhere. Though you do not need to ask for this one: Gapnostic exports your references and reports in standard formats from inside the app, at any time, without spending a credit.
If you are in the UK or EU, these are your rights under data protection law, and you can complain to your national supervisory authority if you think we have handled them badly.
Being straight about the beta
Gapnostic is new. A small number of researchers are using it by invitation while we find out what is wrong with it.
We back up the server, but this is early software and you should keep your own copy of anything you would be upset to lose. That is not us disclaiming responsibility. It is what we would tell a friend.
If something changes
If we ever change how any of this works, we will email everyone with a workspace before it takes effect, not update this page quietly and hope nobody notices.
Asking us something
A person reads this address: ruhu@gapnostic.com. If anything above is unclear or you think we have got something wrong, please say so.