# CiteMap: Documentation Infrastructure for AI Agents

## Core Value Proposition
CiteMap transforms standard developer documentation into machine-readable "Answer Surfaces". 
It solves the problem of "Invisible Documentation" where LLMs (ChatGPT, Claude, Perplexity) cannot parse heavy HTML/JS sites effectively.

## The Problem: Signal-to-Noise Ratio
- Standard websites have ~5% Signal Ratio (mostly HTML/CSS noise).
- AI Agents have limited context windows (128k - 200k tokens).
- Result: Documentation is truncated or ignored, leading to hallucinations.

## The Solution: Parallel Layer
CiteMap deploys a silent, secondary layer alongside your human documentation:
1. **The Miner**: Scrapes and analyzes your content.
2. **The Generator**: Creates high-density Markdown context files.
3. **The Protocol**: Serves these files via /llms.txt and /citemap.xml.

## Key Features
- **Docs-to-Markdown Pipeline**: Auto-converts HTML to optimized context.
- **Visibility Validator**: Audits your site's noise ratio.
- **Analytics**: Tracks AI agent visits and intent.

## Pricing
- **Open Source ($0)**: Self-hosted CLI with unlimited ingestion.
- **Managed ($199/mo)**: Cloud hosting, CD/CI, and analytics dashboard.
- **Enterprise (Custom)**: SLA guarantees and dedicated solutions engineering.

## Technology Stack
- Next.js (Frontend/Protocol)
- Playwright (Miner)
- Pinecone (Vector Memory)
- OpenAI (Semantic Analysis)

Contact: partnerships@citemap.io