Developers

Regulation-as-code.
Install in 30 seconds.

42 interpreted obligations across 7 regulatory frameworks — EU AI Act, ISO 42001, Colorado AI Act, NIST AI RMF, HIPAA, SOX, FTC. Queryable as JSON. Cross-referenced. Cited back to source text.

Free tier available. No credit card to start. Get an API key →

Three ways to use the engine.

All three hit the same underlying rules engine. Same 42 obligations. Same ethics guardrails. Same audit trail.

Claude Code, Claude Desktop, Cursor

One command configures the MCP server across every detected client. Your AI assistant can then cite obligations while you build.

terminal
npx @governmyai/setup

Node / TypeScript SDK

Embed in your service. Query obligations, gate CI/CD, wrap agent tool-use. TypeScript types ship with compile-time enforcement of the human-review flag.

npm install
npm install @governmyai/sdk

REST API & webhooks

Any language. HMAC-signed webhooks with timestamp signatures fire when the engine flags an obligation needing human review.

endpoint
POST /api/rules/obligations

When to use which package.

If you want to... Use this
Ask Claude / Cursor governance questions while coding @governmyai/mcp-server (or just run @governmyai/setup)
Embed compliance checks in your Node / TS app @governmyai/sdk
Wrap Anthropic's tool-use loop @governmyai/sdk-anthropic
Instrument a LangChain agent @governmyai/sdk-langchain
Use the rules engine from any other language /api/rules/* REST endpoints
Receive compliance events as they fire HMAC-signed webhooks (Governmy-Signature header)

We deliberately don't return "compliant: true."

The engine never tells you your AI system is compliant or non-compliant — those are determinations only a qualified human reviewer can make. EU AI Act Article 14 explicitly prohibits delegating that decision to an automated system. Every response carries a humanReviewRequired flag. Obligations marked cannotBeAutoSatisfied always block. This is structural, not configurable.

Seven frameworks. 42 interpreted obligations.

Each rule is machine-readable JSON, cross-referenced to related obligations in other frameworks, and cited back to the source text.

EU AI Act

Arts. 6, 9, 10, 11, 13, 14, 15, 43, 50, 72, 73. Includes Annex III category classification.

ISO 42001

Clauses 4-10 plus Annex A controls A.5, A.6, A.7, A.8, A.9, A.10.

Colorado AI Act (SB205)

Scope, developer obligations, deployer obligations, consumer rights & disclosures, algorithmic discrimination.

NIST AI RMF

Govern, Map, Measure, Manage functions and their core categories.

HIPAA (AI)

AI decision support, PHI in training data, privacy rule authorization, security rule for AI.

SOX & FTC (AI)

Internal controls over financial AI, Section 5 deception & unfairness, endorsement disclosure, ECOA algorithmic fairness.

Ready in 30 seconds.

Grab an API key, pick your path, ship compliance-aware AI.