MCP Tools

The MCP server exposes MidLyr’s core API capabilities as agent-callable tools.

browseRegulatoryLibrary

Use this to find relevant compliance-source documents.

Inputs include:

• query
• category
• authorities
• jurisdictions
• cursor

Returns Regulation Objects.

readRegulatoryDocument

Use this after browsing to fetch a single document’s metadata, total size, and table of contents. Body text is served by readRegulatoryContent.

Inputs include:

• id

Returns a Regulation Detail Object.

readRegulatoryContent

Use this to fetch a document’s plain-text content by character offset. The response always includes the full Regulation Detail Object under details.

Inputs include:

• id
• offset
• limit

Returns a Regulation Content response (text, offset, limit, totalBytes, hasMore, details).

startComplianceScreening

Use this when the agent needs to screen text for compliance risk.

Input:

{
  "content": {
    "type": "text",
    "text": "we can't provide a provisional dispute credit until the investigation is completed"
  },
  "scenario": "dispute"
}

scenario is required and selects the screening context. Valid values: marketing_asset, dispute, debt_collection, complaint, generic.

Returns:

{
  "id": "job_01HY3K4M7QR9VP2N8WYJF5GTB"
}

getJob

Use this to retrieve async job status and typed result payloads.

Input:

{
  "id": "job_01HY3K4M7QR9VP2N8WYJF5GTB"
}

Returns a universal job envelope. When a screening job succeeds, type is screen_analysis and result is { type, riskScore, findings } — a result type discriminator, risk score, and list of compliance findings with regulatory citations.

See Tool Schemas for full tool schemas.