Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.complyhat.ai/llms.txt

Use this file to discover all available pages before exploring further.

ComplyHat is an MCP server that generates compliance documentation for your AI and ML systems. Connect it to any MCP-capable host — Claude Code, Claude Desktop, Codex, or any custom client — and your host agent can register models, run bias tests, generate regulatory reports, and maintain a compliance wiki without leaving its native interface.

Quickstart

Add ComplyHat to your MCP host and make your first tool call in minutes.

Authentication

Learn how OAuth 2.1 and Dynamic Client Registration work from your perspective.

Tool Reference

All 14 entity tools and their modes, with request and response examples.

Supported Frameworks

Nine regulatory frameworks including EU AI Act, SR 11-7, NAIC, and more.

​
What ComplyHat does

ComplyHat connects to your MCP host and gives it 14 compliance tools. Your host agent decides what to call and when; ComplyHat executes deterministic bias, drift, explainability, and adversarial tests and returns structured citations — never synthesized prose or hallucinated findings. Every prose field is prefixed [EXTRACTED], [INFERRED], or [AMBIGUOUS] so reviewers and auditors can see exactly what the document actually says versus what was inferred.

​
Supported hosts

ComplyHat works with any MCP-capable host that speaks streamable-HTTP:
  • Claude Code — claude mcp add complyhat https://complyhat.ai/api/mcp
  • Claude Desktop / Cowork — paste the URL into claude_desktop_config.json
  • Codex Desktop / CLI — add the URL to your Codex MCP config
  • claude.ai connectors — add via Settings → Connectors
  • Any custom MCP client — point at https://complyhat.ai/api/mcp

​
Supported frameworks

EU AI Act

Annex IV technical file for high-risk systems. Deadline: 2 August 2026.

SR 11-7

Federal Reserve model risk management documentation and validation evidence.

NAIC Model Bulletin

Insurance AI governance, testing, and consumer-protection sections.

NIST AI RMF

Govern, Map, Measure, Manage attestations for US federal programs.

ISO/IEC 42001

AI management system control evidence for ISO certification audits.

NYC Local Law 144

Annual independent bias-audit report for automated employment decision tools.

Treasury FS AI RMF

Control-evidence mapping for financial institutions running AI in core operations.

Colorado SB 21-169

Risk assessment and consumer-impact disclosures for Colorado insurance and AI.

CMS-0057-F

Clinical-validation and bias-testing artefacts for Medicare/Medicaid AI.

​
How to get started

1

Sign up

Create an account at complyhat.ai. Your first tool call will trigger an OAuth consent screen — sign in and approve to link your workspace.
2

Add the MCP URL to your host

Paste https://complyhat.ai/api/mcp into your host’s MCP configuration. See the quickstart for host-specific instructions.
3

Register your first model

Ask your host to register a model. It will call the models tool with mode: "register" and return a model ID you use in every subsequent call.
4

Generate a compliance report

With a model registered, ask your host to generate an EU AI Act report. ComplyHat returns structured, audit-tagged sections your host can stitch into the final deliverable.