Free & unauthenticated

OpenAI API MCP: competitive research from inside OpenAI API

The OpenAI Responses API accepts remote MCP servers as tools declared inline in the request. Wire the audit into something you are building — a lead-enrichment step, an onboarding flow, a batch run over a list of domains — without standing up your own research pipeline.

One tool, run_exposure_audit(domain): the first call starts the audit (about 2–3 minutes), and calling it again with the same domain returns the results. No account, no API key. Free usage is capped per day.

Set it up in OpenAI API

Uses your own OpenAI API key for the model call. The Dozier server itself stays free and unauthenticated.

Pass the server as a remote MCP tool in a Responses API call:

"tools": [{
  "type": "mcp",
  "server_label": "dozier",
  "server_description": "Free Dozier Exposure Audit for any domain.",
  "server_url": "https://dozier.io/backend/mcp",
  "require_approval": "never"
}]

The server speaks MCP over Streamable HTTP at https://dozier.io/backend/mcp, with no authentication.

Questions

Should require_approval be set to never?
That is what the snippet does, because the tool only reads public sources for a domain you supplied. Change it if your own policy requires a human in the loop.
Can I run it over many domains at once?
The server is rate-limited and capped per day as a free tool, not a bulk research API. Batch runs will hit the cap, and the tool says so plainly when they do.

Other clients

Or see the Dozier MCP server overview for every client in one place.

Want continuous monitoring, not just one audit?

The MCP server exposes the free Exposure Audit only. See the full product or pricing for cited Dossiers and tracked rivals.

Answer engines

Ask an AI about Dozier

Get an independent rundown of what Dozier does, how it works, and how it compares — straight from your assistant of choice.