Privacy Policy
Last updated: June 2026
What We Collect
- Account data: email address, display name, username, optional bio and profile picture.
- Subset entries: the text items you submit to each subset (e.g., your 7 Wikipedia articles).
- Vector embeddings: numerical representations of your entries, computed locally on our server. These are 384-dimensional floating-point vectors — not human-readable.
- Interaction data: which subsets you've joined, which users you've pinned.
What We Do NOT Collect
- Gender, age, location, or any demographic information.
- Browsing history, cookies from third parties, or tracking pixels.
- Data from any external API or social media account.
Cookies
We use a single authentication cookie (hg_uid) to keep you signed in. This is a strictly necessary cookie — it contains only a signed session identifier and is not used for tracking, analytics, or advertising. No third-party cookies are set.
Legal Basis for Processing
We process your data on the basis of consent, which you provide by creating an account and submitting entries. The authentication cookie is processed under "strictly necessary" grounds. You may withdraw consent at any time by requesting account deletion.
Age Requirement
You must be at least 16 years old to use this service. We do not knowingly collect data from anyone under 16. If we learn that a user is under 16, we will delete their account and data.
How We Process Data
Your text entries are converted to vector embeddings using a local ONNX model running on our server. No data is sent to any third-party AI service (no OpenAI, no Google, no cloud inference). All computation happens on our own hardware.
K-Anonymity
Matching results are withheld until at least 49 people have submitted entries to a given subset. This K-anonymity threshold ensures that your approximate position in the vector space cannot be trivially isolated.
What Other Users See
Other users can see: your display name, username, bio, profile picture, and which subsets you've joined. They can see an approximate similarity score if both of you are in an active subset. They cannot see your raw text entries.
Data Storage
All data is stored in SQLite databases on our self-hosted server. Vector data is stored in a separate sqlite-vec database. Profile pictures are stored as WebP files on disk.
Data Deletion and Export
You may request deletion of your account and all associated data, or export of your data in a machine-readable format, by contacting us at the address below. Upon deletion, your entries, vectors, and profile data will be permanently removed.
Contact
For privacy-related requests (deletion, export, questions), contact: [email protected]
Changes
We may update this policy. Material changes will be noted on this page with an updated date.