What the server keeps
You are pointing a real client at a service we run, so here is what it holds. Every line on this page was read out of the server's own source rather than written for it.
Trust and data
What the server keeps
What not to send
Use synthetic data. Do not point this at production, and do not send us personal data, customer data, credentials or secrets. It is a fault-injection sandbox, not a staging environment.
Isolation
Records are scoped to the account that made them, in the list and in the live panel. You see your own and nobody else's. Requests that arrive without a valid token are recorded without an account, so nobody sees those in the live panel at all: only the seven-day rule clears them.
Cost
Free to use.
What is recorded
One row per request: the method, the phase, the status, the duration, whether you cancelled, the protocol revision your client declared, the scenario that was active and the email you signed in with. Parameters are summarised, not captured. On a tool call we keep the tool name and drop the arguments. On a legacy connection we keep the protocol version and your client's name and version. Anything else is kept only if it is 512 characters or fewer.
Headers
Six header values are kept by allowlist: accept, content-type, content-length, host, user-agent and mcp-protocol-version. Every other value is replaced with [redacted], so the value of your Authorization header is never recorded. Header names are kept, values are not.
Retention
Records are deleted after seven days. Pull anything you want to keep into your own tracker.
Deletion
The console has an ERASE EVERYTHING button. It permanently deletes your observations, your saved scenario, your MCP access token and your live sessions and signs you out. It cannot be undone. We hold no account record beyond those rows, so there is nothing else of yours left here. Revoking this app's access to your Google or GitHub account is a separate step, on their side.
Where it runs
Amsterdam. The service and its database run in the EU, and your records are stored there. One exception: feedback you submit in the console goes out through a US email provider and is not covered by the erase button.
Tracking
None. No analytics, no error tracking, no session recording, no third-party scripts. Sign-in talks to Google and GitHub, feedback goes out by email and that is the whole list.
Sign-in
We ask Google and GitHub for your email address and nothing else. No name, no avatar. We do not store their tokens.
Rate limits
120 requests a minute on the MCP endpoint, per IP. Over that you get a 429. Console actions are limited separately.
Marketing
Your sign-in email is not used for marketing.
Source code
The server is open source under the MIT License, at github.com/cakewalk-security/cakewalk-mcp-test-kitchen. The hosted copy this page describes is deployed from that repository, so you can check what it records against the code.
The TypeScript reference client is separate. It is public on GitHub under the MIT License.
Read next
The privacy notice covers the same ground in the language a legal review needs. How to use the MCP Test Kitchen walks through what actually gets written to the log while you test, and the FAQ answers the shorter versions of these questions.
Ready to break something?
Sign in with Google or GitHub, copy your access token, then select a scenario and point a client at the endpoint.