Data Processing Agreement
Version 1.0 · effective 2026-11-12
This agreement covers personal data that may be contained in your requests to the serp.hilgard.cz API, for example a search query built from what your user typed. It sets out what we do with that data, where, for how long, and who else is involved, as required by Article 28 of the GDPR. It applies to every customer from the first request. If you need a signed copy, write to jan.hilgard@gmail.com and you get it countersigned.
1. Parties
Processor: Jan Hilgard, sole trader registered in the Czech Trade Register, company ID (IČO) 67808522, Kostelecká 756, 250 65 Bášť, Czech Republic. Contact: jan.hilgard@gmail.com.
Controller: the customer who holds an API key for serp.hilgard.cz (“you”). Where you act as a processor for your own customers, we are your sub-processor and this agreement applies in the same way.
2. What is processed and why
Subject and purpose: running Google searches on your instruction, each API request being one instruction, and returning the results: organic results, ads, People Also Ask, the AI Overview with its sources and, when you ask for it, the raw HTML of the page. We use the data for nothing else: no profiling, no training of models, no resale, no combining with other data.
Data: the search query and its parameters (country, language, location or coordinates, device), and the results returned. Data subjects: your end users, and people named in a query. We do not ask for any personal data in a request; what a query contains is up to you. If queries may contain sensitive data, use zero data retention (section 4).
Duration: as long as you use the API. On ending, nothing needs to be returned; the data expires as listed in section 3, and on request we delete what can be deleted earlier.
3. What we keep, and for how long
For a normal API key:
| Data | Where | How long |
|---|---|---|
| Query, parameters, status and duration of each request | Operational logs on our servers | Up to 7 days, then deleted by log rotation |
| Results of a query | In-memory caches of our servers, shared by identical queries | Up to 1 hour, never written to disk |
| The results page last loaded | Memory of the browser that ran the search, never written to disk | Until that browser runs its next search |
| Copies of results pages Google served unreadable or without an AI Overview, for troubleshooting | Files on our search servers | Up to 7 days |
| Async searches (POST /v1/searches): query, parameters, result | Database on our API server | Until 24 hours after the search ends |
| Raw HTML requested as html: "link" | Files on our API server, readable only with your key | 24 hours |
| Real addresses behind Google result links, with the result title and site | Memory and files of our servers, not linked to your account or query | Up to 14 days |
| Number and cost of requests per day | Database on our API server, no query text | While the account exists, and as long as accounting law requires |
4. Zero data retention
On request we switch an API key to zero data retention. For such a key, queries and results are processed only in the working memory of our servers while the request runs and are never written down: not to a log, a cache, a queue, a file or a database, on any of our servers. The browser page that ran the search is cleared when it ends. Every search and link resolve response for such a key carries the header X-Data-Retention: none, and GET /v1/usage shows zero_data_retention: true.
Features that need storage are therefore not available for such a key, and the API says so with a 400: async searches (POST /v1/searches) and raw HTML as a link. Raw HTML inside the response works. What is kept:
| Data | How long |
|---|---|
| Time, key id, HTTP status, number of pages and duration of each request, without the query, place or coordinates | Up to 7 days in the operational logs |
| Number and cost of requests per day | While the account exists, for billing |
5. Where the data is processed
Only in the European Union: on our servers in Germany (Hetzner, see section 6) and on our own servers in the Czech Republic. Traffic between our servers runs over an encrypted VPN. We do not transfer your data outside the EU or the EEA.
Google: to answer a request we send its query to Google Search; that is the service you ask for. Searches run in browser sessions that are not signed in to any Google account, and nothing that identifies you or your users is sent with them. Google is not our sub-processor: what Google does with a search it receives is governed by Google’s own terms, and Google may process it outside the EU. Do not put into a query anything you would not send to Google.
6. Sub-processors
| Sub-processor | Address | What for |
|---|---|---|
| Hetzner Online GmbH | Industriestr. 25, 91710 Gunzenhausen, Germany | Hosting of the API server and a search server (data centre in Germany) |
You authorise these sub-processors. We tell you by e-mail at least 30 days before adding or replacing one; you may object, and if we cannot address the objection, you may stop using the API and get back unused prepaid credit. Every sub-processor is bound by data protection obligations no weaker than this agreement.
Payments are handled by Polar (polar.sh) as merchant of record, and account e-mails are sent through Seznam.cz. They see your account and billing details, never the content of your API requests.
7. Security
- The API is served only over HTTPS (TLS). Searches reach Google over HTTPS end to end.
- API keys are stored only as SHA-256 hashes; a key is shown once, when it is issued.
- Servers are administered only over SSH with keys; internal services listen only on the local host or the VPN.
- Data files are readable only by the service account; stored items are deleted on the schedule in section 3.
- Only the processor himself has access to the servers and the data; he is bound to confidentiality.
8. Our obligations
- We process the data only on your documented instructions, given by your API requests and this agreement, unless EU or Czech law requires otherwise; we will tell you if we think an instruction breaks the law.
- We help you, as far as we reasonably can, to answer requests from data subjects and to meet your obligations under Articles 32 to 36 GDPR. As we keep almost nothing, and nothing linked to a person, this help is mostly confirming what section 3 says.
- We tell you about a personal data breach affecting your data without undue delay, and at the latest within 48 hours of learning of it, with what we know.
- We give you the information needed to show that this agreement is kept, and allow audits by you or an auditor you appoint, after reasonable notice and at your cost.
9. Final provisions
This agreement is part of your agreement to use the API. If they conflict on personal data, this agreement prevails. It is governed by Czech law. We may update it; a change that reduces your protection takes effect only after 30 days’ notice by e-mail. Questions: jan.hilgard@gmail.com.