Every device-intelligence call processes personal data. IP addresses, browser characteristics and the identifiers derived from them all describe a person, however indirectly. So when a compliance team reviews a fraud tool, two questions come up: where is this processed, and who can reach the data? A data residency rule answers the first one for you, and it is one of the few requirements that can rule out a managed service no matter how it is configured.
This post covers what residency rules actually ask for, which of them a managed service can satisfy, where it can’t (we’ll be specific about Prynt), and what self-hosting involves when it is the only option.
What residency rules actually ask for
“Data residency” covers several different requirements, and it matters which one you have:
- Statutory localization. Some laws require that certain data, often financial, health or government data, is stored or processed inside the country. Transfers out may be restricted or banned.
- Transfer rules. Laws like GDPR don’t require local processing, but they regulate transfers to other countries. Transfers are allowed with a mechanism such as an adequacy decision or standard contractual clauses, documented in a DPA. See cross-border data transfer and fraud signals.
- Contractual residency. Enterprise customers often require in-region processing in their contracts, whatever the law says.
- Sector rules and regulator expectations. Banking, health and public-sector work frequently add their own localization or outsourcing rules.
The related idea, data sovereignty, asks whose laws govern the data. Data processed in one country by a company subject to another country’s laws can be exposed to those laws even if it never moves. That is a question about the provider, not only the data center.
The practical split: a transfer rule can be met by a managed service in another country, with the right contract. A localization rule can only be met by processing in the required place, by the vendor or by you.
What a managed service can and can’t change
A managed service can reduce how much personal data it receives, who can read the results and how long data lives. It can’t change where processing happens unless it runs in that region. Here is how that breaks down for Prynt today:
| Requirement | What Prynt offers today |
|---|---|
| Processing location | United States only, on servers Prynt operates; no EU or other region |
| Transfer documentation | A DPA, signable on Enterprise |
| Sub-processor disclosure | Listed on the trust page: Cloudflare (DNS, CDN, TLS; sees request metadata including IPs), Stripe (billing), a transactional email provider |
| Retention | Plan windows (30, 90 or 365 days, or as agreed), deleted daily; backups expire after 7 and 14 days |
| Erasure and access | Erase by visitorId or by your linkedId; export visitors and events |
| IP storage | Stored in full in the shared cloud; truncated or not stored on dedicated Enterprise deployments |
| Consent | GPC honored by default; requireConsent holds identification until consent |
If your requirement is a transfer rule, that table is the material for your review. If it is a localization rule for anywhere other than the US, Prynt does not meet it today, and configuration won’t change that.
Reducing what crosses the border
When a managed service is allowed, the next question is how little you can send. Three habits matter most.
Send opaque ids, not identities. linkedId should be your internal user id, not an email or phone number, and tag should name the action, not the person. See linking accounts after signup.
Keep the verdict out of the browser. Have the page send only a requestId, and fetch the result from your server with the secret key. With sealed results the browser receives only an encrypted blob that your server decrypts with a key you hold:
<script src="https://api.pryntid.com/cdn/prynt.umd.js"></script>
<script type="module">
const prynt = await Prynt.load({
apiKey: 'pk_live_…',
endpoint: 'https://api.pryntid.com',
requireConsent: true, // nothing is identified until the visitor agrees
sealedResult: true, // the browser receives an encrypted blob, not the verdict
});
// Call this from your consent banner's "accept" handler.
prynt.setConsent(true);
const { requestId, sealedResult } = await prynt.identify({ tag: { action: 'signup' } });
</script>
Sealing protects the result from the browser, extensions and tampering in transit. It does not keep anything from Prynt, which computes the result; it is not a residency control. Sealed results explained has the server-side decryption.
Keep your own copy small. Your database is in scope too. Store the decision you made and the reason, not every signal, and expire it on your own schedule. Data minimization for fraud signals covers what to keep.
When self-hosting is the answer
Self-host when a localization rule, a regulator or a contract requires device signals to be processed inside a boundary your vendor doesn’t offer, or when the system must run without outbound calls. In those cases the choice is between building on open-source fingerprinting libraries and writing the server, matching and detection around them, or buying from a vendor that sells an on-premise edition.
A self-hosted deployment that actually meets a residency rule keeps every stage inside the boundary:
- The collection script is served from your own domain, not a vendor CDN.
- The collection and scoring API runs in the required region, so raw signals never leave it.
- Storage, retention and deletion happen in your database, on your schedule.
- Keys, logs and backups stay inside your environment too. Backups in another region break the rule as surely as the primary database would.
And it comes with the work a managed service would otherwise do: deploying and patching the service, an on-call owner for a check that sits on your signup and login paths, scaling under attack traffic, and keeping bot and tampering detection current as evasion tools change. Plan for that before you commit; detection that isn’t maintained degrades without anyone noticing. Build vs buy device intelligence helps size it.
Location is only one requirement
Wherever processing happens, the rest of the privacy program still applies:
- Lawful basis. Fraud prevention is often a legitimate interest under GDPR, but document it. See GDPR and device fingerprinting.
- Minimization. Collect and keep only what the fraud decision needs.
- Retention limits. Define how long identifiers and events live, and delete automatically.
- Transparency and rights. Disclose the processing and be able to answer access and erasure requests.
- Regional rules. For the US, see CCPA fingerprinting compliance; for Brazil, LGPD and device fingerprinting; for the general question, is device fingerprinting legal.
Frequently asked questions
What is data residency versus data sovereignty?
Data residency is where data is physically stored and processed. Data sovereignty is whose laws govern that data. A rule about residency is met by processing in the right place; a sovereignty concern is about which company and which country’s laws can reach the data, which processing location alone does not settle.
Can a managed fraud detection service meet data residency requirements?
Only if it processes in the region your rule names. A managed service can minimize what it receives, encrypt results for your server, delete on a schedule and sign a DPA, but none of that changes where processing happens. Prynt processes in the United States only today.
Does self-hosting fingerprinting make me automatically GDPR compliant?
No. Self-hosting removes a third-party processor and keeps data where you put it, but you still need a lawful basis, minimization, retention limits, transparency and a way to honor access and erasure requests.
Start by naming the rule you’re under: localization, transfer, contract or sector. A transfer rule can usually be met with a managed service and a DPA; a localization rule outside the vendor’s region can’t. For the full list of what you control with a managed service, see self-hosted vs SaaS fraud detection, and for the broader trade-offs, why teams self-host fraud detection.
Try it free
Prynt is device intelligence with a free plan: 5,000 identifications a month, visitor IDs with confidence and bot detection — no credit card. Upgrade to Pro for VPN, proxy, tampering and behavioral signals.