HUMMBL is an AI governance infrastructure company. It ships governance as a Python library — stdlib-only, open source — rather than as a SaaS dashboard. The primitives (Circuit Breaker, Kill Switch, Governance Bus, Delegation Tokens, Delegation Context) give engineering teams runtime-enforced governance they can audit, own, and extend.

How much does HUMMBL cost?

The hummbl-governance SDK is free and open source (Apache 2.0 license). The AI Governance Arbiter audit, consulting engagements, and HUMMBL Certified Adapter assessments are paid services. See hummbl.io/pricing.html for current pricing.

What AI frameworks and regulations does HUMMBL support?

HUMMBL primitives are actively aligning to NIST AI RMF, ISO 42001, EU AI Act, and Colorado AI Act requirements. The governance crosswalk at hummbl.io/crosswalk.html shows which controls each primitive addresses across these frameworks.

HUMMBL - Governance Infrastructure for AI-Native Teams

Try in Claude Desktop in 30 seconds

Install the HUMMBL MCP server to give Claude access to 120+ mental models for structured reasoning.

npm install -g @hummbl/mcp-server

Add to claude_desktop_config.json:

claude_desktop_config.json

{
  "mcpServers": {
    "hummbl": {
      "command": "npx",
      "args": ["@hummbl/mcp-server"]
    }
  }
}

Restart Claude Desktop and ask: "What mental models should I use for this problem?"

The Governance Tuple

Every governed AI decision produces a triple: CONTRACT (what was authorized), DCT (who authorized it, cryptographically signed), and EVIDENCE (what actually happened). This is the atomic record that makes AI governance auditable.

CONTRACT

Permitted operations, resources, caveats, expiry

DCT

HMAC-signed delegation binding issuer → subject

EVIDENCE

Actual operations, resources accessed, cost, output

governance_tuple.py

from hummbl_governance import DelegationTokenManager, BusWriter, CostGovernor

# CONTRACT: define what the agent may do
contract = {
    "ops_allowed": ["read", "summarize"],
    "resources": ["docs/*"],
    "expiry": "2026-12-31T23:59:59Z"
}

# DCT: cryptographically bind the delegation
dtm = DelegationTokenManager()
token = dtm.issue(
    issuer="orchestrator",
    subject="summarizer-agent",
    operations=contract["ops_allowed"],
    resources=contract["resources"]
)

# EVIDENCE: audit what actually happened
bus = BusWriter("governance.jsonl")
bus.append(agent="summarizer-agent", action="read", resource="docs/report.md")
bus.append(agent="summarizer-agent", action="summarize", resource="docs/report.md")

Read the Paper → OWASP 10/10 Coverage →

Cognitive Layer: Base120 Mental Models

The reasoning substrate that operates over governance primitives. 120 validated mental models across 6 transformation types.

Perspective

Frame and anchor viewpoint. Change how you see the problem.

Inversion

Flip to see what avoiding failure reveals. Think backwards.

Composition

Combine for emergent properties. Build complexity from simplicity.

Decomposition

Break complexity down. Understand the parts.

Recursion

Apply patterns at multiple scales. Iterate and refine.

Systems

Understand governing rules. See the meta-patterns.

Need governance for your agents?

Free Readiness Assessment

30-minute call. I'll review your agent architecture against EU AI Act Annex III requirements and tell you where the gaps are.

Learn More → View latest Supadata snapshot (JSON)

Try It Now

Describe a problem you're facing and get recommended mental models instantly. No signup required.

Recommended Models

Explore All 120 Models →

Claude Desktop Setup

Install the MCP server to give Claude direct access to all 120 models.

Step 1: Install

terminal

$ npm install -g @hummbl/mcp-server

Step 2: Configure

Add to claude_desktop_config.json

claude_desktop_config.json

{
  "mcpServers": {
    "hummbl": {
      "command": "npx",
      "args": ["@hummbl/mcp-server"]
    }
  }
}

Step 3: Available Tools

select_model Select appropriate mental models for a problem
apply_transformation Apply a transformation type to reframe a problem (P, IN, CO, DE, RE, SY)
analyze_problem Comprehensive analysis using multiple mental models

Step 4: Available Resources

models://all All 120 mental models
models://by-transformation Models grouped by transformation type
relationships://all Validated relationships between mental models
transformations://overview Overview of the 6 transformations

Step 5: Available Prompts

problem_decomposition Systematic 6-transformation problem analysis framework

Use Cases

HUMMBL provides structured reasoning infrastructure for AI systems. Use it via API, Claude Desktop MCP integration, or direct framework integration:

Agent Frameworks

LangChain, CrewAI, AutoGPT, Semantic Kernel - invoke mental models programmatically to guide agent reasoning.

Claude Desktop

MCP server gives Claude direct access to all 120 models. Search, retrieve, and apply cognitive frameworks in real-time.

Custom Applications

REST API for decision support tools, prompt engineering systems, or reasoning-enhanced workflows. No framework required.

Prompt Engineering

Retrieve relevant mental models based on problem context, then inject into system prompts for better LLM outputs.

Knowledge Systems

Build cognitive enhancement tools, decision support systems, or structured thinking applications on top of Base120.

Research & Education

Access validated mental models programmatically for cognitive science research, educational tools, or training systems.

Book a call to see how Base120 fits your use case.

Schedule Discovery Call →