Privacy and data handling.
We collect as little as we can get away with. This page says exactly what is stored, why, how long, and what you can ask us to do about it.
Analytics
There is no Google Analytics here, no Plausible, no third-party tag of any kind, and no tracking cookies — which is why you were never asked to accept any.
We keep a self-hosted request log on our own server: the path requested, the referrer, the user-agent string, a country code from Cloudflare's edge header, and a salted hash of the IP address. The hash lets us count visitors without storing who they are. We keep this mainly to see which AI crawlers — ClaudeBot, GPTBot, PerplexityBot and the rest — read the site. Rows are deleted after 90 days.
Security logging
One exception to the hashing rule: when a request is a recognisable attack —
probing for /.env, SQL injection, path traversal, a vulnerability
scanner's user-agent — we store the raw IP address alongside
the reason it was flagged. We need it to block the source and, where the traffic
is persistent, to file an abuse report with the network that hosts it. Legal
basis: legitimate interest in keeping the service available (Art. 6(1)(f) GDPR).
These rows are deleted on the same 90-day schedule.
Newsletter
If you subscribe to the dispatch we store your email address, the time you subscribed, and a hashed form of your IP address for abuse prevention. Signup is single opt-in: your address is added when you submit the form, and we immediately send a welcome email that contains an unsubscribe link.
Every dispatch carries a one-click unsubscribe link and the standard
List-Unsubscribe headers, so your mail client can unsubscribe you
without opening anything. Unsubscribing stops all sending immediately; write to
us if you also want the record deleted outright.
Contact
When you email [email protected] or any other Empero address, we keep the message for as long as we need it to deal with your request and to meet record-keeping obligations.
Contributor applications
When you apply through /hiring we store what you enter in the form (name, email, area, country or time zone, links, your answers and your talent-pool choice), your skill-task answers if you take the task, a salted hash of your IP address for rate limiting, and our review notes and decisions. Basis: steps taken at your request before entering into a collaboration (Art. 6(1)(b) GDPR) and, for the talent pool, your consent (Art. 6(1)(a)), which you can withdraw at any time.
AI-assisted review. Before your written answers are reviewed, we remove email addresses, phone numbers and postal addresses from them. The remaining text, including your name, is processed by an external AI service provider, which may process it in the USA, to score it against a fixed rubric and summarize it for us. The provider may not use it for training. A strong result can move your application on to the skill task automatically; the AI never rejects an application. A person reviews every application it does not advance and makes every rejection and invitation, so no decision about you is based solely on automated processing (Art. 22 GDPR). Transfers to the USA rely on the EU-U.S. Data Privacy Framework or standard contractual clauses. You can ask us which provider we use.
Retention. Applications are deleted automatically 6 months after the last activity on them, or 24 months if you joined the talent pool. You can ask us to delete yours at any time.
Comments and the community
To comment on posts, vote on comments, like posts or save them, you sign in with your Empero account at accounts.empero.org. That service tells this website your account's ID, its email address, whether the address is confirmed, and the account's name. We store these with the name you choose to show, your comments, your votes on comments, the posts you like and save, the reports you send, and when you last signed in. Your comments and the name next to them are public, and so are the numbers of likes and votes. Who liked or voted is not shown, and your email address never is.
Before a comment is shown, a filter on our own server checks it for spam and abusive language: comments it flags wait for a moderator, and slurs and threats are not accepted. To deal with spam and abuse we also keep a salted hash of your IP address with each comment for 30 days. Moderators can hide or delete comments and stop an account from commenting. Basis: the service you asked for (Art. 6(1)(b) GDPR) and our legitimate interest in keeping the comments free of spam and abuse (Art. 6(1)(f)). You can delete your comments, likes and saved posts at any time on your community page; deleting your Empero account deletes them as well. accounts.empero.org has its own privacy notice.
Contributor portal
If you have an account on the contributor portal at /portal, we store your name,
email address and role, the profile details you add (title, bio, photo, time zone), what you post
in the forum, the calendar events you create or are invited to, the articles you write and their
review notes, and when you last signed in. We use this to run the portal and to send the emails it
needs: sign-in links, review requests and decisions, mentions and meeting invitations, and the daily
digest if you switch it on. Basis: our collaboration with you (Art. 6(1)(b) GDPR) and our legitimate
interest in running the lab's internal tools (Art. 6(1)(f)).
Other members see your name, role, profile and what you post; the visitor numbers on the portal's dashboard are counts only and never show who a visitor is. To have your account and what you posted deleted, write to [email protected].
Cookies
The public site sets no cookies unless you sign in to comment: then one
strictly-necessary cookie keeps you signed in, and another lasts for the fifteen minutes the
sign-in itself may take. Neither is used for tracking. The admin area at /admin and the contributor portal at /portal each set one strictly-necessary session cookie for signed-in members; it is not used for tracking and never reaches visitors.
Infrastructure
The site runs on our own server in Europe, behind a Cloudflare tunnel that terminates TLS. Cloudflare processes request metadata (including IP addresses) as part of delivering the connection. Model weights and code are hosted by Hugging Face and GitHub; visiting those links takes you to their platforms, under their privacy policies.
Your rights
Under the GDPR you have the right to:
- Access the personal data we hold about you
- Rectify inaccurate data
- Erase data (the “right to be forgotten”)
- Restrict or object to processing
- Data portability
- Lodge a complaint with a supervisory authority
To exercise any of these, write to [email protected] — for a newsletter record, sending from the subscribed address is the fastest route.
Controller
Empero — independent AI research lab, Germany. Data-protection questions: [email protected].