IntelligentOperations.ai Docs
IntelligentOperations.ai is an AI-Operating System that lets you execute parallel AI agent workflows via pre-built prompt libraries and context briefs. This documentation covers the MCP server, REST API, and integration standards.
The website is the MCP server. Every page has a /md twin. The /api/mcp endpoint implements the full MCP HTTP transport protocol and is available to any connected AI client.
Core concepts
Context Briefs are structured JSON documents containing your company context — name, industry, voice, products, goals. Every agent in every library reads the same brief before executing. Write once, use everywhere.
Prompt Libraries are named collections of agents organized by use case. Each library contains 3–20 agents that run in parallel when executed. Libraries are categorized by domain: Website, Marketing, Sales, Legal, Finance, Engineering, and more.
Agents are individual LLM workers with a specific system prompt, model assignment, and bounded task. Each agent is stateless — it reads the context brief, executes its task, returns a result, and terminates.
Executions are parallel agent runs. When you execute a library, all agents run simultaneously via Promise.all(). Results stream back in real time via Supabase Realtime.
Architecture
intelligentoperations.ai/
├── /api/mcp ← MCP HTTP server (GET: capabilities, POST: JSON-RPC)
├── /api/agents ← Agent execution API (POST: run, GET: results)
├── /platform ← Platform UI (auth, briefs, libraries, results)
├── /[slug]/md ← Markdown twin for every page
├── /llms.txt ← AI discovery manifest
└── /.well-known/ ← mcp.json, apple-app-site-association