Stateless and Read-Only by Design
Does Bassethound store the domains I analyze?
Yes. Results are stored against your account, the way any hosted service stores your work. The calls themselves stay independent, so nothing you looked up earlier changes a later answer, and the data is public information about a domain.
Best move: treat each call as independent, and assume your results are stored the way any hosted service stores your work.
Why it works: statelessness here is a property of the request boundary, not a promise to forget. Each call carries its own domain and profile, and nothing from a previous call changes what you get back. Your results are stored against your account.
Key takeaways
- Your results are stored against your account, the way any hosted service stores your work.
- Each analysis record holds the requesting IP and your account identity, and is retained indefinitely.
- No sniff_domain call shares state with any other, so nothing you looked up earlier shapes a later answer.
- Results are cached by domain and profile for up to 24 hours, so a repeat call may return a cached dossier.
- The server is read-only against your side: it fetches the target domain, not your data.
What does stateless mean here?
Statelessness is a property of the request boundary. Each sniff_domain call carries everything it needs: the domain and the profile. Call it twice for the same domain and the second call does not consult the first. There is no session object, no cursor, no per-user state that changes your result.
Stateless describes what flows between calls, not what gets saved. Your results are saved and tied to your account.
Read-only is the other half, and it holds without qualification. Bassethound reads the target domain: its HTML, DNS, TLS, and headers. It does not read or write anything on your side. The only input from you is the two arguments.
What is stored?
The domain you submitted, the options you passed, the resulting dossier, your account identity, and the IP the request came from. The privacy policy is the authoritative statement on retention.
Why does independence still matter for an agent?
Reproducibility. A call is a function of the domain and the profile at crawl time, so the same inputs give you the same shape of output. No hidden per-user state changes what comes back, which means you can cache results on your side, diff them over time, and reason about them.
It also keeps autonomous loops predictable. Bassethound ships as a remote MCP endpoint any host can connect to (Claude, Cursor, ChatGPT), and the worst case of a runaway agent is wasted rate-limited calls, not a corrupted server-side profile that skews later answers.
What independence does not buy you is correlation across calls. To compare a domain today against last quarter, you hold that history. The deep profile’s Wayback first-seen history is the exception, and it reads a public archive.
Bassethound perspective
The moat is fusion at request time: five correlated layers in one call, on a domain, computed when you ask. Storing your results is account management, not the product. We do not sell the data or build a profile of you from it.
Sources
- Model Context Protocol specification: https://modelcontextprotocol.io/specification
- MDN, an overview of HTTP (stateless protocol): https://developer.mozilla.org/en-US/docs/Web/HTTP/Overview
- IETF RFC 9110, HTTP Semantics: https://datatracker.ietf.org/doc/html/rfc9110
Frequently asked questions
Does Bassethound keep a database of the domains I look up?
Yes. Results are stored against your account. Calls stay independent of one another, so nothing you looked up earlier changes a later answer.
Does Bassethound record my IP address?
Yes. It is kept in standard server logs, and it is also stored with each analysis record, which is retained indefinitely. On the MCP endpoint your daily allowance counts against the account you signed in as, not your address.
Does signing in start storing my queries?
Signing in is what ties results to your account. The MCP endpoint requires a sign-in.
Is the returned dossier cached and reused?
Yes, for up to 24 hours. Results are cached by domain and profile, so a repeat call for the same domain can return the cached dossier rather than recrawling.
Can my agent call sniff_domain at volume without building a profile somewhere?
Fanning out stores one result per call against your account, the same as any hosted API. The dossiers are public information about the domains you asked about.