Effective date: 16 July 2026 · Last updated: 9 October 2026
Operator: Alexandr Maximov, IČO 88788814, Jeseniova 1196, 130 00 Prague, Czech Republic ("usefulapi", "we", "us"). Contact: [email protected].
usefulapi provides hosted Model Context Protocol (MCP) servers that let AI clients (e.g. Claude, Cursor) access third-party services (such as Linear) on your behalf. This policy explains what we collect, why, and your choices. It covers usefulapi.io and the *.usefulapi.io MCP endpoints. We are the data controller for the personal data described here.
Account & identity. When you connect a server, you authenticate to the underlying third-party service via OAuth (or, for some servers, an API token you provide). We receive and store a credential (OAuth access/refresh token, or your API token) so the server can call that service for you. We derive a per-user identifier from it to key your usage and billing. We do not receive your third-party account password.
Email login (servers without a credential). Servers for public data sources need no credential; you log in with your email address instead, and we send a 6-digit code from [email protected]. The address is used only to send that code. We do not store it: we keep only a one-way hash of it, which is your account identifier for usage and billing.
Usage metering. We record counts and timestamps of tool calls per user and per server to enforce plan limits and bill you. We do not need, and do not retain, the *contents* of your tool calls to meter them. To avoid counting a repeated call twice, we keep a one-way hash of the tool arguments for about 10 seconds.
Product telemetry. For each tool call we also record the server, the tool name, the outcome (for example success, or the class of a vendor error), the duration, the size of the result and the name of the type of AI client (for example Claude or Cursor). This record has no user id, no tool arguments and no results. We keep it for 90 days to find tools that fail.
Feature requests. When your AI client uses a server's <prefix>_request_feature tool, we store the text it sends with the server name, the type of AI client and the name of the tool that you tried (if any), not with your account, for up to 90 days, to plan new tools. Do not put personal data in such a request. To allow at most 5 requests per day, we count them per account for 24 hours; this count does not contain the text.
Request contents (transient). To fulfil a tool call, the server passes your request to the third-party API and returns the response to your AI client. This data is processed in transit to complete your request and is not stored by us beyond what is technically necessary to serve the response, except transient logs (see §4). Our servers process tool arguments and results in memory only, to complete each call. We do not log or keep their contents, and no person at usefulapi reads them. A server can filter, reformat or shorten a result, or add a usage notice to it, before it reaches your AI client. The only exception is a feature request: we keep and read the text that you send with a <prefix>_request_feature tool.
Billing. Payments are handled by Stripe. We do not receive or store full card numbers; Stripe does. We store your Stripe customer/subscription identifiers and plan status.
Operational logs. Our infrastructure (Cloudflare) processes IP address, request metadata, and error diagnostics for security, abuse prevention, and reliability.
We process personal data to: perform our contract with you (providing the service, billing); pursue our legitimate interests (securing the service, preventing abuse, debugging); and
comply with legal obligations (accounting/tax). Where we ever rely on consent, you may withdraw it at any time.
We do not sell your data, and we do not use your data or request contents to train machine learning models.
We rely on: Cloudflare, Inc. (hosting, DNS, storage of encrypted tokens), Stripe (payments), and the third-party service you connect (e.g. Linear) — your requests go to that provider under its own privacy policy. A current list is available on request at [email protected].
Under the GDPR you may request access, correction, deletion, export (portability), restriction, or object to processing, and withdraw consent where applicable. Contact [email protected]; we respond within one month. You can disconnect a server at any time, which revokes our stored credential for that service. You also have the right to lodge a complaint with the Czech Data Protection Authority (Úřad pro ochranu osobních údajů, uoou.gov.cz).
Credentials are stored encrypted and scoped per user; webhook and payment flows are signature- verified. No method of transmission or storage is perfectly secure, but we take measures appropriate to the sensitivity of the data.
We and our sub-processors may process data in the European Union and the United States (Stripe and Cloudflare are US-based). Where data leaves the EEA, transfers rely on appropriate safeguards (e.g. Standard Contractual Clauses and/or the EU-US Data Privacy Framework).
The service is not directed to individuals under 16, and we do not knowingly collect their personal data.
We may update this policy; material changes will be posted here with a new effective date.
Alexandr Maximov — [email protected] — Jeseniova 1196, 130 00 Prague, Czech Republic.