Skip to main content
The Human Design API is designed for programmatic consumption: structured JSON, machine-readable error codes, a consistent response envelope, and a comprehensive OpenAPI spec. This makes it straightforward to integrate as a tool for AI agents and LLM workflows.

Why This API Works Well with AI

  • Consistent structure — every response has the same envelope (success, errorCode, data)
  • Machine-readable errors — agents can branch on errorCode without parsing text
  • camelCase fields — no special casing surprises for JSON parsers
  • OpenAPI spec — import directly into tool-building frameworks
  • Deterministic — same birth data always produces the same chart

llms.txt

This documentation site serves an auto-generated llms.txt file:
Feed this to your LLM for full API context. A more comprehensive llms-full.txt with the complete documentation content is also available at /llms-full.txt.

OpenAI Function Calling

Define the API as a tool in the OpenAI function calling format:
When the LLM invokes this tool, make the API call:

Claude Tool Use

Define as an Anthropic tool:

Coordinates Endpoint for Agents

If your agent already has latitude and longitude (e.g., from a prior geocoding step), use the coordinates endpoint to skip the geocoding key entirely:
This returns a full ChartResult without requiring an HD-Geocode-Key header.

No-Code Platforms (Zapier, Make, n8n)

Use an HTTP request module with these settings: Map dynamic fields from your workflow into the body. Parse data from the response for the chart result.

Error Handling for Agents

Always check success and branch on errorCode:
See the full Error Codes reference.