# Privacy Policy

> Last updated 17 November 2026.

This policy explains what personal data Moatlight (“we”, “us”) handles at moatlight.tech, and what is made public. The scan itself reads public pages of a domain, not personal data. The personal data is small: an email address if you ask for a full report, a hashed network address for rate limits, and a few cookies. Scores and reports about domains are public, and that is covered in its own section.

## Who we are

Moatlight runs the free agent-readiness scan at moatlight.tech and the crawler MoatlightBot. We decide why and how the data below is used, so we are the controller. Write to us at [ceo@moatlight.tech](mailto:ceo@moatlight.tech).

## Scores and leaderboards are public

Anyone can scan any domain. When a scan finishes, its result is stored and shown at /report/<domain> to anyone who has the link. This is the core of the service, so please read it before you scan:

- The free part of a report is open to everyone, with no login: the domain, the score, the grade, the layer scores, the three biggest fixes, crawl readiness and token cost. A report is public even if the person who asked for the scan is not the owner of the domain.
- The full report (every check with its evidence, fix prompts, generated files) opens after someone enters an email address. The email address is never shown on a report, a leaderboard or anywhere else public.
- We plan to publish category leaderboards that rank vendors by score, with the stage that coding agents reached in our test runs. A leaderboard entry shows the domain, the score and grade, the rules version and hash it was scored with, the dates, and a link to the report. Test runs by agents are labelled “simulated” and kept apart from the score.
- Before a vendor is listed, we email the result to a contact we found for that company, and give a 7-day right to reply. A reply is published next to the entry, in full or with a link.
- We may publish the underlying leaderboard data as a downloadable dataset with a licence. It would hold domains, scores and check results, not emails or hashed addresses.
- Reports are currently marked noindex, so search engines are asked not to list them. This may change when a leaderboard goes live, and the public pages may still be copied, cached or archived by others. Deleting something on our side does not remove those copies.

A score belongs to a domain, which is usually a company, not a person. If a domain is a personal site, a score about it can still be personal data, and the rights below apply to it. Domain owners can ask us to correct a result, add a reply or re-run a scan. See “Your choices” and the [Terms](https://moatlight.tech/terms#public-scores).

## What we collect

- The domain you scan, and what the scan finds on that domain’s public pages and endpoints: the results of each check, the pages fetched, timings and errors. This describes a site, not you.
- A rate-limit identity: a salted SHA-256 hash of your IP address (for IPv6, of its /64 prefix), stored with the scan and in the counters that enforce the limits (10 scans a minute, 30 a day). We do not store the raw address in our database. Our hosting provider sees the raw address as part of serving the request, and keeps ordinary server logs.
- If you ask for a full report: your email address, the domain, a record that you agreed to be contacted, facts from the scan about the company behind the domain (score, grade, whether it has an MCP server, llms.txt, an OpenAPI file or OAuth), and, if present, the first-touch attribution described below. We note whether our confirmation email was sent and when you confirmed it.
- Email delivery data: that a confirmation email was sent to you. We do not track opens or clicks in the email.
- Product analytics events, if analytics is switched on: “scan started”, “report viewed”, “email captured” and “email confirmed”, with the domain, score and grade. The identifier is the hashed address above, never your IP address or email, and no person profile is created.
- Messages you send us by email.

We do not ask for your name, do not run advertising or cross-site tracking, and do not sell personal data. We do not use your data to train AI models.

## Cookies

- ml_attr: set by your browser on your first visit, for 30 days. It holds UTM parameters, the referring website’s host name and the first page you landed on, so we can tell which launch or link brought a lead. It is read by us only when you enter your email, and is then stored with that lead.
- ml_unlock: set when you enter your email in the report gate, for 30 days, HttpOnly. It is a signed token that proves you have asked for the full report, so the report opens on later visits. It holds your email address in signed (not encrypted) form and is only ever sent to us.

Both are used to run the service you asked for and to measure which channels work. We set no advertising cookies. If a law in your country requires consent for them, tell us and we will switch the attribution cookie off.

## Why we use it, and the legal basis

- To run a scan and show you the result, and to enforce rate limits: our legitimate interest in providing and protecting the service (GDPR Article 6(1)(f)), and performing what you asked for (6(1)(b)).
- To open the full report and send you a confirmation link: performing what you asked for (6(1)(b)), and your consent when you submit the form (6(1)(a)). You can withdraw consent at any time.
- To write to you once about what the scan found: our legitimate interest in telling a company about a problem it asked to look at, and your consent in the form. We write once, and stop the moment you ask.
- To contact the vendors we list on a leaderboard, and to publish public scores: our legitimate interest in publishing a fair, checkable benchmark, with the right to reply described above.
- To measure which channels bring people and to see where the funnel breaks: our legitimate interest in improving the service. The data is not tied to your identity.
- To keep the service secure, stop abuse and meet legal duties.

## Who handles the data

We use these companies to run the service. They process data on our instructions:

- Vercel: hosts the website and the scan API, and sees request data including IP addresses.
- Supabase: hosts our Postgres database (scans, results, leads, rate-limit counters).
- Railway: runs the scan worker and the render service. They handle domains and public page content, not visitor data.
- Resend: sends the confirmation email, so it sees your address and the message.
- PostHog: receives the analytics events above, if switched on.
- Anthropic: when we run simulated agent tasks against a company’s public site, the public page content and the task go to Anthropic’s API. No visitor data is sent.

We also share a lead’s email and company facts with our own team to follow up. We do not sell or rent personal data, and we do not “share” it for cross-context behavioural advertising. We disclose data if a law or a valid legal request requires it, or in a sale or reorganisation of the business, with notice to you.

These providers may process data outside your country, including in the United States. Where we send personal data from the EU, UK or Switzerland to a country without an adequacy decision, we rely on the provider’s standard contractual clauses or the EU–US Data Privacy Framework. Ask us for a copy of the safeguards.

## How long we keep it

- Scans, check results and reports: kept as long as the report or a leaderboard that cites it is up, so an old score can be traced to the rules that made it. Old scans of a domain stay as history, and only the latest is shown at the report address.
- Leads (email, domain, company facts, attribution): until you ask us to delete them, or until 24 months after the last contact, whichever comes first.
- Rate-limit counters: deleted automatically after their window, which is a minute or a day.
- Hashed rate-limit identity stored with a scan: kept with the scan. It cannot be turned back into an address, but it can be linked to repeat scans from one connection.
- Cookies: 30 days. Server logs at our hosting provider: as long as that provider’s own policy says.
- Emails you send us: as long as needed to answer, and then for up to 24 months.

## Your choices and rights

Write to [ceo@moatlight.tech](mailto:ceo@moatlight.tech) from the address concerned, or reply to any message from us. We answer within 30 days. You can ask us to:

- delete your email address (replying to any message from us is enough, and we do it without questions);
- give you a copy of the personal data we hold about you, correct it, or limit or stop its use;
- stop contacting you, or stop using your data for our legitimate interests;
- correct, add a reply to, re-run or discuss a public score about your domain (see the [Terms](https://moatlight.tech/terms#public-scores)).

If you are in the EU, the UK or Switzerland, you also have the right to data portability, to withdraw consent and to complain to your data protection authority. If you live in California or another US state with a privacy law, you have the rights of access, deletion, correction and opt-out of sale or sharing. We do not sell or share personal data, so there is nothing to opt out of, and we do not treat you differently for using these rights. You can use an authorised agent to ask on your behalf, and we may need to confirm your identity first.

You can also delete the cookies in your browser at any time. The full report then locks again on that browser until you use the confirmation link.

## If you run a scanned site

MoatlightBot reads the public pages of a domain when someone scans it. It makes read-only requests and obeys robots.txt. To stop it fetching beyond your homepage, disallow MoatlightBot in robots.txt. See the [MoatlightBot page](https://moatlight.tech/bot) for what it fetches and its user agent. The request logs on your side will show the scan.

## Security and children

Our database is reachable only from our servers. It is not exposed through a public data API, and its tables have row-level security enabled with no public access. Connections are encrypted. No system is perfectly secure, and if a breach affects you we will tell you and the authorities where the law requires it.

The service is for people building developer products and is not meant for children under 16. We do not knowingly collect their data. If you think we have, write to [ceo@moatlight.tech](mailto:ceo@moatlight.tech).

## Changes and contact

We will change this policy when the service changes, for example when a leaderboard goes live. The date at the top says when it was last updated, and we post material changes on this page before they apply. Questions: [ceo@moatlight.tech](mailto:ceo@moatlight.tech).
