# YakaAI Documentation Hub - Full AI Integration Specification

This document provides a comprehensive guide for AI agents to discover, search, and extract technical documentation from the YakaAI Documentation Hub.

## 🔗 Entry Points
- **Production API Base**: `https://yakaai.com/api`
- **Staging API Base**: `https://master.universal-ai-cheatsheet-hub.pages.dev/api`
- **OpenAPI Spec**: `https://yakaai.com/openapi.json`
- **Web Hub**: `https://yakaai.com/`

## 🛠 Available Tools & Endpoints

### 1. Repository Discovery (`/repos`)
Use this endpoint to find libraries and frameworks. It is optimized for keyword search and metadata filtering.
- **AI Recommendation**: When looking for a library, use `q=<keyword>` and `sort=velocity` to get the most popular and active results first.
- **Output**: Returns JSON with `capabilities` (`hasSkills`, `hasMcp`, `hasOpenApi`), `doc_chars`, and `doc_files`.

### 2. Deep Context Extraction (`/cheat/:owner/:repo`)
This is the primary endpoint for retrieving documentation. It returns a single Markdown file.
- **AI Recommendation**: Always check the first few lines of the output for the `Source:` header. If it says `AI-Optimized (llms.txt)`, the content is already optimized for you.
- **Targeted Section Extraction**: Use `?section=<topic>` to save context window tokens or `?section=toc` for the Table of Contents.

### 3. Native Skills Extraction (`/skills/:owner/:repo`)
Retrieves repo-specific rules (`.cursorrules`, `CLAUDE.md`, `AGENTS.md`) and Model Context Protocol (MCP) server configs.
- **AI Recommendation**: This is mandatory for Zero-Shot development. Before suggesting code changes for a library found in the hub, call this endpoint to follow the author's preferred patterns.

### 4. API Specification Extraction (`/spec/:owner/:repo`)
Retrieves isolated OpenAPI 3.x, Swagger, GraphQL, or Protobuf schemas.

### 5. Dynamic OpenGraph Visual Thumbnails (`/og`)
Generates 1200x630 vector SVG banners at the Edge for social cards and previews.

### 6. Dynamic Topic Catalogues (`/topic/:tag`)
Browse curated hubs: `mcp`, `skills`, `openapi`, `react`, `nextjs`, `hono`, `tailwind`, `database`, `orm`, `auth`, `docker`, `rust`, `python`, `go`, `ai`, `rag`.

## 📦 Data Schema & Types

### Repository Object
| Field | Type | Description |
|---|---|---|
| `full_name` | string | Unique identifier for `:owner/:repo` routes. |
| `name` | string | Repository name. |
| `description` | string | Repository summary. |
| `language` | string | Primary programming language. |
| `stars` | integer | Total GitHub star count. |
| `topics` | array | Topic tags. |
| `updated_at` | string | ISO timestamp of last update. |
| `capabilities` | object | Flags for `hasSkills`, `hasMcp`, `hasOpenApi`. |

## 🚀 Advanced Integration Patterns

### Auto-Ingestion Workflow
1. Call `/api/repos?q=<library_name>&sort=velocity&limit=1`.
2. Extract `full_name`.
3. Call `/api/skills/<full_name>` to load coding rules and MCP configs.
4. Call `/api/cheat/<full_name>` to load API documentation.
5. Proceed with code generation or task execution.

---
*Document Version: 1.1.0 (2026 Edition)*