# Super Search MCP Server

Unified web search + content extraction MCP server powering search across the IT Pro Partner ecosystem.

- **Production version:** v2.3.0 (19 tools, deployed on Core :8899)
- **Repo version:** v2.0.0 (10 tools — repo behind production)
- **Systemd:** `super-search.service`
- **Framework:** FastMCP 4.0+ (Streamable HTTP)
- **MCP Endpoint:** `http://127.0.0.1:8899/mcp`

## Backend Providers
| Provider | Type |
|----------|------|
| SearXNG | Meta-search |
| Exa | AI search + content extraction |
| Brave | Web search |
| Serper | Google SERP |
| Tavily | AI research |
| Perplexity | AI research |
| DuckDuckGo | Web search |
| Firecrawl | Web scraping |
| OpenCorporates | Company search |
| CourtListener | Legal search |
| Wikipedia | Encyclopedia |

## Tools (v2.3.0 — 19 total)
Core: `web_search`, `web_extract`, `web_extract_document`, `web_extract_batch`, `web_search_premium`, `web_search_news`, `web_search_academic`, `web_search_images`, `web_suggest_queries`, `web_lookup`, `health_check`

Advanced: `web_research_deep` (Exa Agent), `lockdown_mode`, provider-specific search variants, OSINT person search

## Deployment
- **Location:** Core (152.53.192.33) — `/root/docker/super-search/`
- **Recovery:** `systemctl restart super-search`
- **Dependencies:** SearXNG Docker container, provider API keys in `/root/.hermes/.env`
- **Crawl4AI:** companion microservice on :8910 for JavaScript-rendered page extraction

## ⚠️ Repo Staleness
This repository contains v2.0.0 source. Production runs v2.3.0. The production source at `/root/docker/super-search/server.py` needs to be pushed to this repo.