Skip to content

For AI Agents: How to Consume This Documentation¶

This page explains how external AI agents (Custom GPTs, Claude Workspaces, LangChain agents, and other cloud AI systems) can consume the Group Run documentation programmatically.

Machine-Readable Context Files¶

The docs site auto-generates two files on every build:

File	URL	Best For
`llms.txt`	`https://azharameen.github.io/group-run/llms.txt`	Fast ingestion — titles and descriptions only
`llms-full.txt`	`https://azharameen.github.io/group-run/llms-full.txt`	Deep context — full page content

How to Use¶

Option 1: Fetch `llms.txt` (Recommended for most agents)¶

import httpx

context = httpx.get("https://azharameen.github.io/group-run/llms.txt").text
# Pass `context` as system prompt or tool result to your agent

Option 2: Fetch the OpenAPI schema for API-aware agents¶

schema = httpx.get(
    "https://backend-service-601546984807.asia-south1.run.app/openapi.json"
).json()
# Use schema to auto-discover API routes, request/response schemas

Option 3: Direct page fetch for targeted context¶

architecture = httpx.get(
    "https://azharameen.github.io/group-run/architecture/"
).text

Key Architectural Facts¶

Supervisor pattern: All agent calls route through a LangGraph supervisor graph
Config files: config/teams.yaml and config/mcp.json are hot-reloadable without restart
Storage: SQLite for checkpointing + YAML workspaces for idea/artifact storage
Port: Container binds to ${PORT} env var (Cloud Run requirement), defaults to 8080
Health route: GET /api/health → {"status": "ok"}