# Ife Web Intelligence & Data Extraction Engine — Full Manifest

> Complete API Specification, Endpoint Endpoints, Schemas, and Remote Model Context Protocol (MCP) Capabilities.

Base URL: `https://ife.sluxia.com/api/v1`  
Remote MCP Endpoint: `https://ife.sluxia.com/mcp`  
Support Email: `support@ife.sluxia.com`  
Security Email: `security@sluxia.com`

---

## 1. REST API Specification

### Authentication
All production REST API endpoints require an API Key passed in either the `X-API-Key` HTTP header or as a Bearer token:
```
Authorization: Bearer ife_live_...
X-API-Key: ife_live_...
```

### Endpoints Overview

#### 1. POST /api/v1/synthetic-dataset
- **Description**: Generates Q&A instruction pairs in JSONL format for OpenAI, Alpaca, ShareGPT, and DPO fine-tuning.
- **Request Payload**: `{ "url": "string" }`
- **Response Payload**: `{ "source_url": "string", "total_pairs_generated": number, "dataset_health_score": number, "dataset": Array<{ "question": "string", "ideal_answer": "string" }> }`

#### 2. POST /api/v1/security-audit
- **Description**: Scans web pages for Indirect Prompt Injections (hidden CSS/comments targeting AI crawlers) and leaked API key secrets.
- **Request Payload**: `{ "url": "string" }`
- **Response Payload**: `{ "url": "string", "security_score": number, "risk_level": "string", "vulnerabilities": { "leaked_secrets": Array, "prompt_injections": Array } }`

#### 3. POST /api/v1/geo-competitor
- **Description**: Head-to-head AI recommendation audit comparing domain semantic density and citation probability against a competitor.
- **Request Payload**: `{ "your_url": "string", "competitor_url": "string", "topic": "string" }`

#### 4. POST /api/v1/aeo
- **Description**: Analyzes AI search crawler access (GPTBot, ClaudeBot, PerplexityBot) and generates Organization / FAQ JSON-LD schemas.
- **Request Payload**: `{ "url": "string" }`

#### 5. POST /api/v1/rag-chunk
- **Description**: Chunks web content into vector-ready blocks preserving header ancestry (`H1 > H2 > H3`) with exports for Pinecone, Qdrant, Weaviate, and Chroma.
- **Request Payload**: `{ "url": "string", "chunk_size": number, "overlap": number }`

#### 6. POST /api/v1/table-extract
- **Description**: Parses HTML `<table>` elements into structured JSON row objects, GFM Markdown tables, and CSV files.
- **Request Payload**: `{ "url": "string" }`

#### 7. POST /api/v1/clean-article
- **Description**: Reader-mode article distiller that prunes boilerplate navigation and returns clean body markdown, lead summary, and hero images.
- **Request Payload**: `{ "url": "string", "max_length": number }`

#### 8. POST /api/v1/code-extract
- **Description**: Extracts code blocks from technical documentation, classifying programming languages across 12 syntaxes.
- **Request Payload**: `{ "url": "string" }`

#### 9. POST /api/v1/media-extractor
- **Description**: Discovers media assets including HLS/DASH video manifests (`.m3u8`), podcasts, hero images, and alt-text accessibility.
- **Request Payload**: `{ "url": "string" }`

#### 10. POST /api/v1/link-graph
- **Description**: Builds an internal link equity graph categorizing internal routes, external domains, downloadable files, and social profiles.
- **Request Payload**: `{ "url": "string" }`

#### 11. POST /api/v1/llms-generator
- **Description**: Crawls core domain pages to generate standard `/llms.txt` and `/llms-full.txt` developer manifests.
- **Request Payload**: `{ "url": "string", "company_name": "string" }`

#### 12. POST /api/v1/brand-mentions
- **Description**: Monitors online brand mentions across web and news while excluding self domain, returning sentiment breakdown and Share of Voice.
- **Request Payload**: `{ "brand_name": "string", "brand_domain": "string", "recency": "30d" }`

#### 13. POST /api/v1/pdf-to-markdown
- **Description**: Downloads and converts PDF documents into structured Markdown.
- **Request Payload**: `{ "url": "string" }`

#### 14. POST /api/v1/crawl
- **Description**: Crawls target URLs returning multi-format output (markdown, HTML, plain text, links).
- **Request Payload**: `{ "url": "string" }`

#### 15. POST /api/v1/markdown
- **Description**: Converts any webpage or HTML element directly into clean GFM Markdown.
- **Request Payload**: `{ "url": "string" }`

#### 16. POST /api/v1/extract
- **Description**: Extracts structured JSON data using custom CSS selectors and attribute targets.
- **Request Payload**: `{ "url": "string", "selectors": Object }`

#### 17. POST /api/v1/search
- **Description**: Multi-provider web search engine (Bing, DuckDuckGo, Wikipedia) with domain authority classification and snippet proximity highlighting.
- **Request Payload**: `{ "query": "string", "max_results": number }`

#### 18. GET /api/v1/health
- **Description**: Monitors system health, MySQL liveness, memory usage RSS, process uptime, and system health score.

---

## 2. Remote Model Context Protocol (MCP) Server

- **Server Endpoint**: `https://ife.sluxia.com/mcp`
- **Supported Transports**: Streamable HTTP (`POST`) & Server-Sent Events (`GET /mcp`)
- **OAuth 2.1 Metadata**: `https://sluxia.com/.well-known/oauth-protected-resource`
- **Supported Tool Names**:
  1. `brand_mentions`
  2. `search`
  3. `aeo`
  4. `rag_chunk`
  5. `security_audit`
  6. `synthetic_dataset`
  7. `crawl`
  8. `markdown`
  9. `geo_competitor`
  10. `llms_generator`
  11. `pdf_to_markdown`
  12. `table_extract`
  13. `clean_article`
  14. `code_extract`
  15. `media_extractor`
  16. `link_graph`
  17. `extract`