MCP server for AI agents
Let AI assistants screen people and vendors against OIG LEIE, SAM.gov and state Medicaid lists through a remote MCP server, with evidence and provenance on every result.
What you can do
Screening as agent tools
Eight tools cover the API: screen a person, screen an organization, batch screening, batch results, start and stop monitoring, list coverage and read usage.
Nothing to install
A remote server at `https://api.exclia.com/mcp` over Streamable HTTP. Point Claude Code, Cursor, VS Code or Claude Desktop at it and add your API key.
The same rules as the API
Every tool calls the `/v1` API with your key, so rate limits, plan allowances and metering are exactly the API’s. A refusal comes back as a tool error the assistant can explain.
Built for a person in the loop
Results are `clear` or `potential_match` with per-field evidence and the list version checked. The server tells the assistant that a potential match is never a determination about a person.
How to use it
- Step 1
Create an API key
Owners and admins open API keys in the app and create a key for the assistant, separate from your production integration.
- Step 2
Add the server to your client
In Claude Code run `claude mcp add --transport http exclia https://api.exclia.com/mcp` with an `Authorization: Bearer` header. The MCP guide has configs for Cursor, VS Code and Claude Desktop.
- Step 3
Ask in plain language
For example: “Screen the three home-care agencies in this spreadsheet and tell me which lists you checked.” The assistant chooses the tools and summarizes what came back.
- Step 4
Review potential matches yourself
Open each potential match in the app, compare the evidence, and confirm identity with the publishing agency before recording a decision.
- Step 5
Monitor what matters
Ask the assistant to enroll the people and vendors you keep working with, so future list updates are checked without another prompt.
Questions
MCP (Model Context Protocol) is an open standard that lets AI assistants call outside tools. Exclia’s MCP server turns its screening API into tools an assistant can use on your behalf.
Any client that supports remote MCP servers over Streamable HTTP, including Claude Code, Cursor and VS Code. Claude Desktop and other stdio-only clients connect through the mcp-remote bridge.
No. It reports a potential match with the evidence behind it. A person at your organization confirms identity with the publishing agency and records the decision.
No separate charge. Each screening an assistant runs counts toward your plan exactly as an API call does, and the get_usage tool shows where you stand.
The names and identifiers you ask the assistant to screen pass through your AI client, under that provider’s terms. Exclia returns results to the client; it does not share data with the AI provider itself.
See it on your own list
Start a free trial and screen your first records today, or run a free check on one name first.