mcp-omnisearch
A unified search MCP server: one interface to Tavily, Brave, Kagi, Exa, Linkup, Firecrawl and GitHub search
- 358 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The server calls external search services with your keys
- The provider keys are secrets and some services are paid
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/mcp-omnisearchDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The server calls external search services with your keys
- The provider keys are secrets and some services are paid
This is third-party code. Review the repository files before installing.
What it does
The server gives the agent unified access to several search and extraction services through one MCP interface. It covers web search via Tavily, Brave, Kagi and Exa, answer-style ai search, search over GitHub code, repositories and users, and page content extraction via Firecrawl and Tavily. Each service needs its own key; providers without a key are skipped and the rest keep working. It installs via npx or builds from source, with keys set in the MCP configuration as environment variables.
Who it is for. For developers, analysts and writers who need web search and page extraction from inside the agent.
Good fit when
- You need web search from the agent across several providers
- You need to extract page text for further work
- You need search over GitHub code and repositories
Not a fit when
- You have no keys to any search service
- The agent's built-in web tool is enough
Example request
Search recent material on the topic via Tavily and extract the text of the two best pagesMCP
- Transport
- stdio
- Authentication
- key is optional
Environment variables
TAVILY_API_KEYsecret- Tavily key for web search and page extraction
BRAVE_API_KEYsecret- Brave Search key for web search
KAGI_API_KEYsecret- Kagi key for search and answer tools
EXA_API_KEYsecret- Exa key for web and ai search
FIRECRAWL_API_KEYsecret- Firecrawl key for page content extraction
LINKUP_API_KEYsecret- Linkup key for ai search
GITHUB_API_KEYsecret- GitHub token for code, repository and user search
Limitations
The server itself is free, but the search services behind it need their own keys and run on their pricing and limits. The providers are external services, availability from some regions without a VPN is not guaranteed, and some need a paid account. The keys are secrets; keep them out of the config via variable expansion where the client supports it.
How to disable. Remove the mcp-omnisearch entry from your client's MCP configuration and delete the provider keys you set.
FAQ
Do I need keys for all services?
No, set keys only for the providers you need; the rest are skipped automatically.
How is the server installed?
Via npx with the mcp-omnisearch package, or by building from source with pnpm install and pnpm run build.
Discussion
No questions or reviews yet.
Related
MCP serversMCP server for real-time web search: general and vertical search, batch queries and page content extraction
Official Tavily MCP server: real-time web search, page content extraction, site mapping and crawling
Metasearch for agents: one MCP server and HTTP API merge results from many search providers into unified JSON with deduplication.
Brave's official MCP server: web, news, image, video and place search plus extracted page content for model context
Local web search for agents: searches, reads pages and returns compact passages with sources, saving context tokens
Yandex and Google search MCP server
web-compass
An MCP web-search and content-extraction server via Yandex and Google on top of SerpAPI, geared toward RU/CIS results
README in short
The README describes an MCP server that offers unified access to several search services, ai tools and content extraction. It supports Tavily, Brave, Kagi, Exa, Linkup, Firecrawl and GitHub search, each with its own key, and providers without a key are skipped. Installation runs via npx or mcpick, with a build-from-source option. Keys are set in the MCP configuration as environment variables, and the README advises storing them through expansion. The tools web_search, ai_search, github_search and web_extract are documented. MIT licensed.