AI & Integrations

GEOG provides machine-readable resources to help AI assistants, LLM-powered tools, and automated systems understand and integrate with the platform.

llms.txt

Following the llmstxt.org convention, GEOG publishes structured context files for large language models:

These files help AI coding assistants (Cursor, Claude Code, GitHub Copilot, etc.) generate correct API integration code without requiring you to paste documentation into the conversation.

Usage

Point your AI tool at the llms.txt URL, or paste its contents as context:

# In an AI assistant prompt
Read https://geog.dev/llms.txt for GEOG API context

OpenAPI Specification

The API specification is available as an OpenAPI 3.1.0 document:

The OpenAPI spec can be used with:

• API clients — Import into Postman, Insomnia, or HTTPie
• Code generators — Generate typed API clients with openapi-generator or similar tools
• AI assistants — Feed the spec for precise endpoint understanding

AI-Friendly API Patterns

The GEOG API follows conventions that work well with automated and AI-driven integrations:

Places Search

The places search endpoint accepts structured JSON and returns predictable results:

curl -X POST "https://api.geog.dev/v1/places/search" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "coffee shop",
    "location": { "lat": 37.7749, "lng": -122.4194 },
    "radius": 5000,
    "limit": 5
  }'

Vector Tiles

Vector tiles use a simple URL template — ideal for automated map rendering:

GET /v1/tiles/{tileset}/{z}/{x}/{y}.mvt
Authorization: Bearer YOUR_TOKEN

Token Exchange

For automated systems that need short-lived credentials:

curl -X POST "https://api.geog.dev/v1/auth/token" \
  -H "Authorization: Bearer YOUR_LONG_LIVED_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"ttl": 3600, "scope": "tiles:read"}'

Rate Limits for Automated Access

Automated systems should respect rate limits, especially on the free tier:

Monitor the endpoint-specific rate limit headers (X-RateLimit-Tiles-Remaining, X-RateLimit-Places-Remaining) and implement exponential backoff when approaching limits. Paid plans have no rate limits.

For full rate limiting details, see Rate Limiting.

Next Steps

• API Overview — Full API reference
• Authentication — Token types and auth flows
• Best Practices — Integration guidelines