Policies
Privacy and limits
What we never expose, the rate limits, and how to use the service responsibly.
What we never expose
- Email addresses. Not on the public page, not in the API, not in any export. Only a one-way hash appears in URLs.
- Anything the user switched off: the photo, the profile page, or search-engine indexing. Each is a separate switch in the dashboard.
- Your API key, which is stored only as a hash.
Identical answers
For an email that is unknown, unverified or private, the service answers exactly the same way: the default image for avatars and a plain 404 (or null in lookup) for the API. Because a hash is not secret (anyone can compute the hash of an address they know), different answers would reveal which people have an account. Build your integration so it never needs to tell these cases apart.
A user can also hide their whole profile page; then the page, its avatar and its JSON/Markdown/vCard/card exports all return 404.
Limits
| What | Limit |
|---|---|
Avatar URLs (/avatar/{hash}, /{username}/avatar) |
600 requests per minute per IP |
| REST API | 600 requests per minute per API key (configurable per site) |
profiles/lookup |
100 hashes per request |
Public exports (.json, .md, .vcf, .card) |
60 requests per minute per IP |
Avatar size (s) |
16 to 512 pixels |
| Sites per user | 10 |
When you exceed a limit you get 429 with Retry-After.
Caching
Cache politely and your integration stays fast and well inside the limits:
| Data | Suggested cache |
|---|---|
| Avatar image | let the browser/CDN follow Cache-Control (1 hour) and use ETag |
| API profile | up to 1 hour |
| Public JSON | up to 5 minutes (Cache-Control: public, max-age=300) |
If a profile later returns 404, the user hid it or removed it: stop showing the data you stored.
Statistics
Each user can see how often their avatar, profile, exports and API data were used, and on which sites. We store daily counters only: how many times, and the host of the referring site (for example blog.example.com). We do not store IP addresses, browsers, cookies or anything about the visitor, and the owner's own visits, bots and HEAD requests are not counted. Rows older than 90 days are deleted. Only the data of a real photo or profile is counted, never the generic default image, so statistics cannot be used to find out whether an address has an account.
Your account data
Each user can download everything we hold about them (a ZIP with their profile, emails, verified accounts, sites without API keys, statistics and photos) and delete their account from Security in the dashboard. Deleting removes the account, profile, photos, emails, keys and statistics at once; copies in backups disappear when the backup rotates out (about 14 days). Two-step sign-in (an authenticator app, with recovery codes) and a list of active sessions are in the same place, and we email users about a new device, a password or two-step change and a suspension. The full legal texts are the Terms and the Privacy Policy.
Responsible use
- Do not enumerate or guess hashes, or test lists of emails to find who has an account.
- Use a profile's data only to show it to that person's own audience in the context they chose, and respect their later changes by refreshing regularly.
- Respect
X-Robots-Tag: noindex. - Do not call the API from client-side code with your key; proxy through your server.
- Do not resell or republish the data as a bulk dataset.
We may rate-limit or disable keys that break these rules. To report abuse, open an issue in the project's GitHub repository; for a security vulnerability, use the repository's Security tab (private reporting) instead of a public issue.