MCP server

The Model Context Protocol (MCP) server gives any compatible agent — Claude Desktop, Claude Code, Cursor and others — first-class sanctions screening tools. It's a thin wrapper over the Screen100 API, so results are identical.

You'll need an API key. A free key works for single screens.

Get an API key

Tools

screen_entity

Screen one person or company name against all sanctions lists.

batch_screen

Screen up to 100 names in one call (Pro).

get_entity

Fetch the full record for one listed entity by slug.

list_watchlists

List the lists searched, with counts and sources.

Claude Desktop / Claude Code

Add Screen100 to your MCP client config:

claude_desktop_config.json

{
  "mcpServers": {
    "screen100": {
      "command": "npx",
      "args": ["-y", "@screen100/mcp"],
      "env": {
        "SCREEN100_API_KEY": "sk_live_your_key_here"
      }
    }
  }
}

Try it

Once connected, ask your agent to screen a name:

> Screen "Ramzan Kadyrov" against the sanctions lists.

→ screen_entity({ "name": "Ramzan Kadyrov", "type": "individual" })
  result: "match"  (OFAC_SDN · RUSSIA-EO14024 · score 0.97)

Prefer raw HTTP? See the API reference and the OpenAPI spec. Machine-readable product details live at /llms.txt.