The Model Context Protocol (MCP) is an open standard for connecting LLM agents to external tools and data sources. An MCP server exposes a typed catalog of tools that an AI assistant can discover and call — letting the assistant act on real data without bespoke integrations on either side.
Checkr operates two MCP servers, each tuned to a different audience. Pick the one that matches your use case:
Customer MCP
Customers (employers and hiring teams) who want their AI assistant to browse and summarize their own background check reports. Connect from Claude, Cursor, VS Code, or any MCP-compatible client — no credentials to provision, just OAuth Dynamic Client Registration.
Read the Customer MCP guide →
Candidate MCP
Embed an LLM agent in your product that retrieves a candidate's own report on their behalf. Provision OAuth Client Credentials in the Checkr Dashboard, then connect Candidate MCP into your agent so candidates can ask "what's the status of my report?" inside your app.
Read the Candidate MCP guide →
| Customer MCP | Candidate MCP | |
|---|---|---|
| Who chats with the agent | A Checkr customer (employer, hiring team) | A candidate viewing their own report |
| How identity is established | Customer signs in with their Checkr account | Candidate verifies via email OTP per session |
| How credentials are obtained | At runtime via OAuth Dynamic Client Registration | Provisioned once in the Dashboard, per Checkr account |
| Production endpoint | https://mcp.checkr.com/customer/ | https://mcp.checkr.com/candidate/ |
Both servers redact sensitive personally identifiable information (PII) from every response. Staging endpoints and the per-MCP tool catalogs are documented on each integration's own page.