Privacy Policy
Effective 28 October 2026.
This policy explains what personal data Anastasia Klimova, trading as SerpKite, 1732 Oberon Crescent, Mississauga, Ontario L4X 2K8, Canada ("SerpKite", "we") collects when you use serpkite.com, the dashboard at app.serpkite.com and the API at api.serpkite.com, why, how long we keep it, and what rights you have. The short version: we collect what we need to run your account and bill you, we never log your query text (a request that fails or returns no results keeps it for 24 hours so we can debug it), results are cached only briefly under a hashed key, and we don't sell data.
1. Controller and processor
We are the controller for account, billing, website and security data. For personal data contained in the queries you send and the results we return to you, we act as your processor under the Data Processing Addendum.
2. What we collect
| Data | Examples | Why | Kept for |
|---|---|---|---|
| Account | Email, name, avatar URL; linked GitHub or Google account ID and verified email; team membership | Sign-in, account security, service email | Until you delete your account |
| Authentication | Hashed session and sign-in tokens, session IP address and user agent | Keep you signed in, let you review and revoke sessions, detect abuse | Session lifetime (up to 30 days), then deleted |
| API keys | SHA-256 hash of each key, key name and prefix, last-used time | Authenticate API requests | Until you revoke the key or delete your account |
| Usage metadata | Timestamp, endpoint, status code, credits, latency, cache hit, key ID, request ID. Never the query text | Billing, spend caps, usage dashboards, abuse prevention, debugging | 31 days, then deleted |
| Failed-request debug records | For a request that fails (5xx) or returns no results only: its query text or target URL, request options (country, language, location, device, page, filters), status, the steps we tried, account and key ID, request ID | Investigate and fix failures (legitimate interest in running the Service) | 24 hours in an in-memory store, then deleted automatically |
| Credit ledger and orders | Credit grants, charges and refunds; pack purchased, amount, Paddle order ID | Balances, accounting, disputes | Kept after account deletion as an anonymised accounting record, no longer linked to your email or name. Invoices and tax records are kept by Polar, our merchant of record, as the law requires |
| Sign-up checks | A keyed hash (HMAC) of the network you sign in from (your IP address, or its /64 block for IPv6) and of a random device identifier kept in the sk_did cookie. Never the IP address or the identifier itself | Stop one person from claiming the free credits again with another account: a new account whose device or network matches another account's gets no free credits, and its API requests are refused until it buys a pack (legitimate interest in preventing abuse; the Terms allow one account per person). Team invitees are exempt. If you think we got it wrong, write to us and a person will review it | Network hashes 30 days and device hashes 90 days after the last sign-in that used them, also after account deletion; then deleted |
| Deleted-account record | SHA-256 hash of your normalised email address, and the deletion time. Never the address itself | Stop the free sign-up credits from being claimed again by deleting and re-creating an account (legitimate interest in preventing abuse) | Kept after deletion; it can't be turned back into your address |
| Settings | Spend cap, alert thresholds, auto-recharge, webhook URL and signing secret | Provide the features you configure | Until changed or account deletion |
| Team invites | Invitee email, hashed invite token | Let the invitee join your team | Until revoked or the team owner deletes the account; invite links expire after 7 days |
| Playground rate limits | IP address of website playground visitors, as a counter key | Prevent abuse of the no-login demo | 24 hours |
| Support email | Messages you send us | Answer you | as long as needed to resolve your request and any follow-up, then deleted |
| Product analytics | Pages viewed (without query strings), product events, pseudonymous ID | Understand and improve the product | See section 5 |
We do not receive or store card details; payments are processed by Paddle, our merchant of record, under its own privacy policy. Encrypted database backups are kept for 30 days, so deleted data disappears from backups within that time.
3. Search queries and results
- Query text is never logged. It is not written to application or database logs, usage records, analytics or the request log in the dashboard.
- Failed requests. When a request fails (an HTTP 5xx answer) or returns no results, we keep its query text, options and the steps we tried for 24 hours in an in-memory store so we can reproduce and fix the failure, then it is deleted automatically. The query text of successful requests is never stored.
- Result cache. To deduplicate identical requests, each successful result is stored in an in-memory cache (not persisted to disk) for up to 6 hours by default, 30 minutes for news, 1 hour for web pages and AI assistant answers, and 24 hours for autocomplete, then it expires automatically. The entry is stored under a SHA-256 hash of the normalized request (endpoint, query, country, language, location, device, page and filters), so the query text is never stored as a key, and the entry holds the result, not your identity or API key. Entries are shared: an identical request from anyone produces the same key.
-
What
max_agecontrols. A cached result is only returned to you when you sendmax_ageand the entry is younger than it. Withoutmax_age(or with0) you always get a live result. Results are written to the cache either way. Raw HTML (include_html) is never cached. - Batch jobs. The request of a batch job is stored until the job runs and then cleared. Its result is kept for 24 hours so you can fetch it or receive it by webhook, then deleted.
-
Crawls and research tasks. The request of a
/v1/crawlor/v1/researchtask (the start URL or research input, and its options) is stored until the task finishes and then cleared. The result (the pages read, or the report and its sources) and the task's webhook URL are kept for 24 hours so you can fetch it or receive it by webhook, then deleted. - AI features. When a request uses an AI feature (include_answer, /v1/answer, /v1/research, and summary or output_schema on /v1/extract; decision checks on short fetched pages (/v1/webpage, /v1/extract, /v1/crawl, include_content) and on page-monitor changes), we send the query or research text and the text of the result pages involved, never your account details, email or API key to our model providers, Anthropic, PBC and OpenAI, L.L.C., under the providers' commercial API terms, which do not allow training on API data; the providers may keep it for a limited period for abuse and safety monitoring. Without those options, nothing is sent to a model provider except the short decision checks listed. Our side: prompts and page text are never logged by SerpKite; generated answers, summaries and extractions are returned to you and not cached, and research reports are kept like other task results.
- Monitors. A monitor's saved search (query, filters and schedule), its webhook URL and any metadata you attach are stored until you delete the monitor or your account (for a page monitor, the watched URL). To report only new results, it keeps truncated SHA-256 hashes of result links it has already reported (at most 1,000 per monitor; for a page monitor, a hash of the page's content). Each run's new results (or the changed page) are kept in the monitor's run history for 24 hours so you can fetch them, then deleted; the run record (status, counts and credits, no query) is deleted after 30 days.
We don't use your queries or results to train models, and we don't share them with anyone except the subprocessors needed to fetch results (section 6).
4. Legal bases (EEA/UK)
- Contract: account, authentication, API service, billing.
- Legitimate interests: security, fraud and abuse prevention (including free-tier abuse), service improvement, and server-side account events used for product analytics (section 5).
- Legal obligation: tax and accounting records.
- Consent: analytics cookies and browser analytics on the website and dashboard (see section 5 and the Cookie Policy). You can withdraw consent at any time.
5. Analytics
We use PostHog, hosted in the EU, for product analytics.
-
Website and dashboard: only with your consent. A banner asks once, and the choice is stored in the
sk_consentcookie for 12 months across serpkite.com and app.serpkite.com. PostHog is not loaded, and sets noph_*cookies or storage, until you click Accept. If you decline, it never loads. - Privacy signals. If your browser sends Global Privacy Control or Do Not Track, we treat it as Decline and don't show the banner.
- Withdrawing. Use "Cookie settings" in the website footer or on the Cookie Policy, or Settings → Cookies in the dashboard. Analytics stops at once, the pseudonymous ID is reset and PostHog's cookies and storage are deleted.
- Server-side account events (such as sign-up, key creation and purchases) are sent by our backend without any cookie, on the basis of our legitimate interest in understanding how the product is used. You can object by emailing us.
Session recording is disabled. Query strings are removed from every URL before an event is sent, so analytics never contains query text, API keys or invite tokens. PostHog keeps analytics events for up to 12 months.
6. Who we share data with
We use a small set of service providers (subprocessors) for hosting, email, payments, logging, analytics and AI features. The current list, with purposes and locations, is on the Subprocessors page. Proxy providers carry the outbound request to the search engine and therefore see the query text in transit, but not your identity or account. Model providers see the query and page text of requests that use AI features (section 3), also without your identity or account. We don't sell personal data and don't share it for cross-context behavioral advertising. We may disclose data if required by law, to protect our rights, or in a merger or acquisition (with notice to you).
7. International transfers
Some providers are established or process data outside the EEA/UK (see the locations on the Subprocessors page). Where they do, we rely on an adequacy decision (including the EU-US Data Privacy Framework and its UK extension, for providers certified under it) or on the European Commission's Standard Contractual Clauses with the UK Addendum, together with the provider's supplementary security measures.
8. Security
API keys, session tokens and sign-in tokens are stored only as hashes. All traffic is encrypted in transit. Our servers accept no inbound connections except through Cloudflare Tunnel. More on the security page.
9. Your rights
Depending on where you live (including under Canada's PIPEDA, the GDPR, UK GDPR and CCPA/CPRA), you can ask to access, correct, export or delete your personal data, object to or restrict processing, and withdraw consent. You can delete your account yourself from the dashboard; this deletes your account data, keys, settings, ledger and usage records. To exercise other rights, email privacy@serpkite.com. We'll respond within 30 days. You may also complain to your data protection authority, or to the Office of the Privacy Commissioner of Canada.
10. Children
The Service is not directed to children and we don't knowingly collect data from anyone under 16.
11. Changes
We'll post updates here and email account holders about material changes before they take effect.
12. Contact
Privacy questions: privacy@serpkite.com. Postal address: Anastasia Klimova, trading as SerpKite, 1732 Oberon Crescent, Mississauga, Ontario L4X 2K8, Canada.
Last updated: 28 October 2026