MCP Omnisearch

MCP Omnisearch is an MCP server that consolidates multiple search engines, AI answer engines, and web scrapers into a unified interface for LLMs. It connects AI assistants to external search and extraction services, including Tavily, Brave Search, Kagi, Exa AI, GitHub, Linkup, and Firecrawl, using four centralized tools. Software developers, researchers, and AI engineers use MCP Omnisearch to execute targeted web searches, retrieve cited AI research reports, inspect GitHub repositories or code snippets, and crawl or parse web pages directly from their preferred client. Rather than maintaining separate MCP servers for distinct search platforms, MCP Omnisearch allows operators to configure only the API keys they own, automatically skipping unconfigured providers while keeping active services operational. The server supports filtering parameters such as time ranges, search depth, and safe search controls, alongside web extraction features that convert remote documents into structured markdown formats. It also accommodates self-hosted Firecrawl installations and manages output sizes through inline or file-based result modes to preserve context window bandwidth during long research sessions.

Category: Browser & Web Automation

Tags: brave, firecrawl, jina, kagi, perplexity, tavily

Visit MCP Omnisearch

How to install and configure MCP Omnisearch

1. Install the server using MCPick by running npx mcpick add --name mcp-omnisearch --command npx --args=-y,mcp-omnisearch. You can specify clients like Claude Code or target VS Code with --client vscode --scope project. 2. Alternatively, build from source by cloning the repository, running pnpm install, and compiling the code with pnpm run build. 3. Configure your MCP client configuration file by adding mcp-omnisearch under mcpServers pointing to node with argument ["/path/to/mcp-omnisearch/dist/index.js"]. 4. Define required API keys in the env block, such as TAVILY_API_KEY, KAGI_API_KEY, BRAVE_API_KEY, GITHUB_API_KEY, EXA_API_KEY, LINKUP_API_KEY, or FIRECRAWL_API_KEY. Providers without keys will be automatically skipped. 5. Restart your MCP client to load the tools: web_search, ai_search, github_search, and web_extract.

What you can do with MCP Omnisearch

• Performing multi-engine web queries across Tavily, Brave, and Kagi to cross-reference technical documentation and recent software releases. - Generating sourced synthesis answers with Kagi FastGPT, Exa Answer, or Linkup to resolve complex questions without manual link navigation. - Inspecting GitHub codebases and searching repository files directly inside Claude Code or VS Code to debug dependencies. - Extracting and parsing web page contents into clean markdown using Firecrawl or Tavily for downstream context injection. - Running asynchronous deep research workflows via Tavily Research using task identifiers to generate comprehensive investigative reports.

Key facts

• https://github.com/spences10/mcp-omnisearch
• Browser & Web Automation, Web Search & Research
• brave, firecrawl, jina, kagi, perplexity, tavily

Part of MCP Servers

What can MCP Omnisearch do?

MCP Omnisearch consolidates web search, AI synthesis, code search, and content extraction into four MCP tools. It queries providers such as Tavily, Brave, Kagi, Exa, and Linkup, performs GitHub repository and code lookups, and parses web content with Firecrawl.

How do I install MCP Omnisearch?

You can install MCP Omnisearch using the MCPick CLI utility by executing npx mcpick add with the appropriate package arguments, targeting clients like Claude Code or VS Code. You can also build it locally using pnpm install and pnpm run build, then register the built script in your client configuration.

Which MCP clients work with MCP Omnisearch?

MCP Omnisearch works with any Model Context Protocol client that supports stdio server processes, including Claude Code, Claude Desktop, and VS Code. MCPick supports automatic installation into supported environments, or you can register the Node.js execution command manually in your client settings file.

Do I need API keys for every provider in MCP Omnisearch?

No, you do not need keys for every supported service. MCP Omnisearch is designed to skip any provider that lacks a configured environment variable. The remaining providers with valid API keys will continue to work without throwing startup or configuration errors.

Is MCP Omnisearch open source?

Yes, MCP Omnisearch is open source software released under the MIT License. The project repository is hosted publicly on GitHub, where users can review the implementation, inspect Vitest test suites, report issues, and contribute improvements following the project contributing guidelines.