HUMMBL APIs

Governance-Augmented Generation (GAG) APIs

With GAG access to the HUMMBL governance kernel, inject governed mental models, compliance analysis, agent control, and verifiable safety into any AI or agent system - with MCP support, as a drop-in complement to traditional LLM APIs and more.

Get API access API reference

Base URL: https://api.hummbl.io  |  MCP: npm install -g @hummbl/mcp-server  |  Spec: OpenAPI 3.1

HUMMBL Governance-Augmented Generation (GAG) APIs

The core reasoning surface: 120 governed mental models with semantic search, recommendation, and multi-step workflow matching - the building blocks agents call to think with HUMMBL.

Mental Models API

Programmable access to the full Base120 mental model catalog and transformations.

GET /v1/models

Recommendation API

Get the mental models most relevant to a natural-language problem description.

POST /v1/recommend

Semantic Search API

3-tier search (Workers AI + Vectorize, D1 cosine, BM25 fallback) - always returns results.

POST /v1/semantic-search

Workflows API

List, retrieve, and match multi-step reasoning workflows to a problem.

GET /v1/workflows

Validation API

Public text validation and sanitization with PII detection - no API key required.

POST /v1/validate

Compliance & Assessment APIs

Analyze governance evidence against the EU AI Act and capture AI governance assessments with automated nurture and scorecard checkout.

Compliance Analysis API

Score governance evidence against EU AI Act articles; get grade, gaps, and remediation effort.

POST /v1/compliance/analyze

Assessment API

Capture assessment results, trigger nurture email sequence, and initiate scorecard checkout.

POST /assessment/capture

Agent Control Plane APIs

Operate a fleet of agents: heartbeat registration, tiered API key management, and a priority task queue with long-polling. Admin-gated.

Agent Registry API

Agent phone-home heartbeats, fleet status, and tiered API key creation/revocation.

POST /agents/heartbeat

Task Queue API

Enqueue, long-poll, and complete priority tasks for the agent swarm.

POST /tasks/enqueue

Safety & Observability APIs

The operational spine: kill switch control, governance proofs, security event logs, metrics, and funnel analytics. Admin-gated.

Safety API

Inspect and update the kill switch; list recent governance proofs per request.

GET /safety/kill-switch

Security API

Arbiter governance validation (ALLOW/WARN/BLOCK), security event log, and statistics.

POST /security/validate

Observability API

Health, request metrics, recent errors, slow requests, and funnel analytics.

GET /metrics

Resources

HUMMBL APIs are designed to be called from any language or agent framework. These resources get you from scratch to a working integration.

OpenAPI 3.1 Specification

The complete machine-readable spec for every HUMMBL API family and endpoint.

openapi.yaml

API Reference

Consolidated human-readable reference: endpoints, parameters, responses, and examples.

reference.md

MCP Server

Expose HUMMBL mental models and search as Model Context Protocol tools to any agent.

@hummbl/mcp-server

Webhooks

Inbound webhook handlers for Cal.com bookings and Stripe payment events.

POST /webhooks/*