# Data processing agreement

> **Draft, 9 November 2026.** Bracketed items are still to be filled in.

This agreement is between the customer ("you", the controller) and [legal entity, registered address] ("we", the processor). It applies to every plan, Free included, and is part of the [terms](/legal/terms). Signing up accepts it. It meets Article 28 of the GDPR and the UK GDPR.

## 1. What we process

| | |
|---|---|
| Subject matter | Analytics for the sites and apps you install nanolytics on |
| Duration | As long as your organization exists |
| Purpose | Collecting, storing, aggregating and answering queries about events |
| Data subjects | Visitors to and users of your sites and apps |
| Instructions | Your configuration of the service and your use of the dashboard, API and MCP |

## 2. Personal data

**Collected for every event:**
- the page address, with the query cut to campaign tags and the fragment dropped;
- the referrer;
- the window width, as a band;
- the user agent;
- the country, region and city, derived from the IP address;
- a visitor ID: a random ID kept in the visitor's browser, or a daily-salted hash of the IP address and user agent when consent is withdrawn or there is no storage.

**Only if you send it:**
- a user ID, through `identify`;
- event names and properties.

**Only if you turn on an add-on:**
- errors: type, message and stack;
- Web Vitals, with a selector of the element measured;
- outbound links and 404s;
- session recordings.

**The IP address** is used for place and the daily hash, and is not stored with events. Our hosting providers' request logs may hold it for a short time.

**Recordings:**
- Replay is off by default. It records a visitor only after they consent.
- Input values are masked by default. Password, card and one-time-code fields are never recorded.
- Text and media can be masked too.

**Don't send special categories of data.**

## 3. Our obligations

1. **Instructions.** We process the data only on your documented instructions, unless the law requires otherwise. If it does, we tell you first unless the law forbids it.
2. **Confidentiality.** Everyone who can access the data is bound to confidentiality.
3. **Security.** We keep the measures in Annex 1.
4. **Subprocessors.** You authorize those on [subprocessors](/legal/subprocessors). We announce a new one there at least 30 days before it starts. You may object on reasonable grounds, and terminate if we can't resolve it. We impose the same obligations on them and remain responsible for them.
5. **Data subject requests.** We help you answer them, including deleting a visitor's or user's data: email [privacy@nanolytics.dev](mailto:privacy@nanolytics.dev) with their visitor ID or user ID.
6. **Help with compliance.** We help with security, breach notification, impact assessments and prior consultation, as far as our part of the processing allows.
7. **Breaches.** We notify you without undue delay after becoming aware of a breach affecting your data, with what we know at the time.
8. **End of service.** When your organization closes, we delete its personal data when you email [privacy@nanolytics.dev](mailto:privacy@nanolytics.dev), unless the law requires us to keep it. Before then, you can export it through the Query API.
9. **Audits.** We make available the information needed to show compliance, and answer an audit request with it first. If an audit is still needed, we allow one a year, on 30 days' notice, at your cost, by an auditor bound to confidentiality.

## 4. Where data goes

- Event data and recordings are stored in the EU: Cloudflare R2's EU jurisdiction.
- They are processed in the Netherlands, on Railway.
- Requests cross Cloudflare's network on the way in.
- Any transfer outside the EEA is covered by the Standard Contractual Clauses (Commission Decision 2021/914, module 3) or an adequacy decision.

## Annex 1: Security measures

- TLS on every connection. Storage is encrypted at rest by its provider.
- Each recording is encrypted with its own key.
- IP addresses are not stored with events.
- API keys are stored only as hashes.
- When a site lists its allowed hosts, browser hits from other hosts are dropped before they are stored.
- Production secrets are held in a password manager and reach systems only at deploy time.
- Access to production is limited to the people who run the service.
- Every number the service answers with carries a query ID, and queries are logged for 30 days.

## Annex 2: Subprocessors

See [subprocessors](/legal/subprocessors).
