MCP Omnisearch Server

Example Use Cases

• 1

  Perform unified web searches across multiple providers (Tavily, Brave, Kagi, Exa, GitHub) without integrating each API separately.

• 2

  Extract, clean, and summarize web content (pages, PDFs, videos) into LLM-friendly text using tools like Jina Reader, Tavily Extract, and Firecrawl.

• 3

  Quick AI answers, citation-backed responses, or to verify/ground statements against web sources using Perplexity, Kagi enrichment, and Jina grounding.

Description

MCP Omnisearch aggregates a variety of search engines (Tavily, Brave, Kagi, Exa, GitHub) and AI/content tools (Perplexity, Jina AI, Firecrawl) behind the Model Context Protocol to provide unified search, AI responses, content processing, and enrichment capabilities. It supports content extraction, summarization, crawling, site mapping, structured extraction, and fact grounding, with output designed for LLM consumption. The server is flexible about API keys — only providers for which you supply keys are enabled — and can be run locally, in Docker, or with a self-hosted Firecrawl instance. It's intended as a developer tool to give agents consistent access to multiple provider capabilities without integrating each provider separately.

Capabilities(2 total)