XposedOrNot MCP¶
Vendor due diligence asks one question that almost never gets answered before signature: how much of this company's email has already been published. This answers it for a domain, without a key.
Server type: Remote (Streamable HTTP) at https://api.xposedornot.com/mcp
Auth: none. Read-only, no key, no account, no session
Live check: unauthenticated POST initialize returned HTTP 200
(serverInfo: XON_MCP, "XposedOrNot Breach Intelligence", version 2.0.0),
tools/list returned the six tools below, and a live tools/call returned the catalog metrics
Vendor: XposedOrNot, xposedornot.com
Why This Matters for Operators¶
A breach check is normally an individual's errand: type your own address into a free site and see whether your password leaked. The business version is different and much harder. Before you hand a vendor production access, or sign a data-processing agreement, or acquire a company, the exposure of its staff email is evidence about how it handles credentials, and it is evidence that no one assembles because the check is per address and a company has hundreds.
The domain view is the one that changes a decision. If a counterparty's domain appears in dozens of breaches with millions of affected accounts, a security questionnaire answer about "strong password policy" is worth less than it looked. If your own domain appears in a breach you did not know about, the response is a scoped password reset, not a press release.
The server also carries the catalog and the aggregate: which breaches exist, when they were added, what industries they hit, and the largest ones by record count. That makes it usable for monitoring a sector as well as checking a name.
The Six Tools¶
| Tool | What it returns |
|---|---|
check_email_breaches |
Whether one email appears in the index, and the breaches it appears in |
get_breach_analytics |
Full breach history for one email: dates, descriptions, exposure detail |
list_breaches |
The breach catalog, filterable by breached company domain or breach ID |
domain_breach_summary |
Aggregate for a whole domain: number of breaches, affected account count, pastes, and the trend |
get_breach_metrics |
System-wide statistics: totals, breaches per year and industry, the largest breaches |
get_recent_breaches |
The most recently added breaches, newest first, with a summary and a source URL |
domain_breach_summary is the tool that earns the entry. It is the difference between a personal lookup and a diligence surface.
What the Live Call Returned¶
A keyless get_breach_metrics call at the time of this sweep returned a catalog of 794 breaches holding 11,717,004,638 exposed records, a separate paste index, and a last-added timestamp of 2026-10-14. That recency matters more than the total: a breach index that stopped updating two years ago is an archive, and this one added entries this month.
Honest Limits¶
- Absence is not innocence. A domain with no hits is missing from this index, which is not the same as never breached. Indexes depend on disclosure, and most breaches are never disclosed.
- Presence is not compromise. A leaked address is exposure of a credential at a point in time. It says nothing about whether that credential was used, whether it was reused, or whether the account still exists.
- Emails are personal data. Checking your own domain is operations. Checking a named individual's address, or bulk-checking a partner's staff, is processing personal data and belongs inside whatever agreement covers that relationship.
- Read-only by design. There is no write surface, so the credential risk of connecting it is the disclosure of what you queried, not of what it could change.
FAQ¶
How is this different from the OSINT server already in the catalog?¶
OpenOSINT MCP carries a breach check as one of nine tools, and it queries HaveIBeenPwned with an API key the operator supplies, one address at a time. XposedOrNot is the dedicated surface: keyless, free, and with a domain-level aggregate that no per-address check can produce.
Should a breach hit block a contract?¶
No single signal should. What it changes is the follow-up: which accounts were exposed, whether the counterparty knew, and what they did after. Treat it as one line in diligence, weighted by what the vendor is about to receive.
Can I monitor my own domain?¶
Yes, and that is the cheapest use of the server: a scheduled check of the company domain against get_recent_breaches turns breach monitoring into a recurring agent task instead of a news item someone forwards.
Does it cost anything?¶
No. The server is free and read-only, with no key and no account, which is why it can be added to an agent today without a procurement step.
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.