The same checks, as an API and an MCP server
The AdminKit Check API is a planned HTTP API and MCP server that exposes the same DNS, mail, TLS, port, MQTT, and Modbus checks behind the free tool pages as a structured, machine-readable result instead of a web page, meant for developers wiring checks into CI/CD, AI agent builders who need programmatic access, and MSPs scripting checks across many client IPs.
Updated October 21, 2026
The problem
AdminKit’s free tools — DNS lookups, mail authentication checks, DNSBL status, TLS checks, port checks, MQTT and Modbus exposure checks — are all built as web pages for a person to read in a browser. That’s a reasonable starting point, but it doesn’t help a developer who wants to run the same check as part of a CI/CD pipeline, or an AI agent that needs to call a check directly and act on the result.
Running the same check for many hosts today means opening the same page repeatedly by hand, which doesn’t scale for an MSP managing dozens of client networks or a team that wants checks to run automatically on every deploy.
There’s also no structured output today. A web page is easy for a person to read but not something code can reliably parse for a pass/fail outcome.
How it works
The plan is for the same checks already running behind the free tool pages to be exposed through an HTTP API and, separately, through an MCP server, so either ordinary code or an AI agent can call them directly. A caller specifies the check and target, AdminKit runs it, and the response comes back as a structured result: pass or fail, the evidence for that outcome, and a fix hint where one applies.
Who it’s for
- Developers who want to run AdminKit’s checks as part of CI/CD or internal tooling.
- AI agent builders who want an agent to call a check directly through MCP rather than scraping a web page.
- MSPs who manage checks across many client IPs and want to script them instead of clicking through pages one at a time.
What exists today, what’s planned
| Capability | Status |
|---|---|
| Every check available as a free web page (see tools) | Free today |
| Preview of the planned OpenAPI contract (/developers/openapi/) | Preview today |
| Preview of the planned MCP server (/developers/mcp/) | Preview today |
| Live, authenticated HTTP API | Early access |
| Live MCP server | Early access |
| Higher-quota authenticated access | In development |
The preview pages describe the planned contract; they are not a live, callable API yet. Until the API ships, the only way to run a check is through its tool page.
Frequently asked questions
Is the Check API live yet?
No. It is early access / in development. Today every check already exists as a free web page under /tools/, built for a person reading a browser, not for code.
What will a result look like?
The plan is a structured response with a pass/fail outcome, the evidence behind it, and a fix hint, the same information the tool pages show today, just in a format code or an agent can parse directly.
Will there really be an MCP server too?
Yes, that's the plan: the same checks exposed both as a conventional HTTP API and as an MCP server, so an AI agent can call them directly using MCP.
Who is this for?
Developers wiring checks into CI/CD or internal tools, AI agent builders who want programmatic access to checks, and MSPs who want to script checks across many client IPs instead of clicking through pages one at a time.
Is there a public contract for this yet?
Preview pages for the planned OpenAPI contract and MCP server already exist under /developers/openapi/ and /developers/mcp/. The live, authenticated API itself has not shipped.
Want early access to Check API and MCP Server for Developers | AdminKit?
Tell us about your setup and we will reach out when it is ready for you — no spam, no newsletter.