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.

A compliance report in ComplyHat is a structured document assembled from your model’s registration metadata, the latest completed bias tests, and the framework template you select. Every prose field in the output is prefixed with an audit tag — [EXTRACTED], [INFERRED], or [AMBIGUOUS] — so your compliance team can see at a glance which claims come directly from your data versus which were inferred. Your host agent stitches these structured sections into the final deliverable; ComplyHat never synthesizes prose internally.
Reports are drafts for human review. Have your compliance team sign off before submitting any report to a regulator.

​
Generate a report

Call the reports tool with mode: "generate", supplying the model_ids array of one or more registered model IDs and the template_slug for the regulation you are targeting. 
{
  "tool": "reports",
  "arguments": {
    "mode": "generate",
    "model_ids": ["mdl_01j9z..."],
    "template_slug": "eu-ai-act"
  }
}
The response includes an id field (rpt_...) and a content object containing structured, audit-tagged sections. The status field will be completed when the report is ready. You can optionally pass a custom name and a configuration object for template-level overrides: 
{
  "tool": "reports",
  "arguments": {
    "mode": "generate",
    "model_ids": ["mdl_01j9z..."],
    "template_slug": "sr-11-7",
    "name": "Q2 2026 Model Risk Review — fraud-detector-v2"
  }
}

​
Supported framework slugs

Pass one of these values as the template_slug argument:
SlugRegulation
eu-ai-actEU AI Act (Regulation (EU) 2024/1689) — Annex IV technical file
sr-11-7Federal Reserve / OCC SR 11-7 — model risk management
naic-model-bulletinNAIC Model Bulletin on the Use of AI by Insurers
nist-ai-rmfNIST AI Risk Management Framework 1.0
iso-42001ISO/IEC 42001:2023 — AI Management System
nyc-ll144NYC Local Law 144 — Automated Employment Decision Tools
treasury-fs-ai-rmfUS Treasury Financial Services Sector AI RMF
colorado-sb21-169Colorado SB 21-169 / SB 24-205 AI Act
cms-0057-fCMS-0057-F — Medicare/Medicaid AI prior-authorization
For a full description of what each framework template renders, see supported frameworks.

​
Retrieve a report

Once generated, retrieve the structured content or metadata at any time with mode: "get". 
{
  "tool": "reports",
  "arguments": {
    "mode": "get",
    "report_id": "rpt_01k2..."
  }
}
To list all reports in your organization, use mode: "list": 
{
  "tool": "reports",
  "arguments": {
    "mode": "list"
  }
}

​
Download the PDF

ComplyHat can generate a signed URL for the report PDF. Call mode: "get_pdf_url" with the report ID: 
{
  "tool": "reports",
  "arguments": {
    "mode": "get_pdf_url",
    "report_id": "rpt_01k2..."
  }
}
The response contains a time-limited signed URL you can download directly or share with your legal and compliance team.

​
Audit-tagged citations explained

Every prose field in a report is prefixed with one of three tags:
TagMeaning
[EXTRACTED]The statement is drawn directly from your model metadata, bias test results, or wiki pages — the source is verifiable.
[INFERRED]The statement is derived from available data but requires an inferential step. Review carefully before signing off.
[AMBIGUOUS]ComplyHat could not find supporting evidence. This field needs manual input from your team.
When your host agent synthesizes the final report prose, it uses these tags to decide which sections to present as-is and which to flag for reviewer attention.
Reports stamp the framework template version they were built against. When a regulator amends the underlying rule, the version stamp lets your counsel verify currency at a glance.
Next steps: Run bias tests before generating reports to ensure [EXTRACTED] test results appear in the output, or explore all reports tool modes in the tool reference.