ResearchMCP
by riii111
ResearchMCP is a research tool that combines the Model Context Protocol (MCP) with search engines like Brave Search and Tavily Search to provide research capabilities similar to ChatGPT's DeepResearch. It allows integration with Claude Desktop for web search functionality.
Last updated: N/A
What is ResearchMCP?
ResearchMCP is a tool that integrates the Model Context Protocol (MCP) with web search engines to enhance research capabilities within applications like Claude Desktop. It aims to provide a research experience comparable to ChatGPT's DeepResearch by leveraging Brave Search and Tavily Search.
How to use ResearchMCP?
To use ResearchMCP, you can either set it up locally or within a Docker container. Locally, you need to install Deno, clone the repository, set API keys for Brave Search and Tavily Search, and then run the appropriate make command for either the HTTP server or the MCP server. For containerized deployment, install Docker and Docker Compose, set environment variables in a .env file, and then build and run the container using make d-build and make d-up.
Key features of ResearchMCP
Web Search via Brave Search API
Full MCP Compliance
Search Result Caching
Integration with Claude Desktop
Use cases of ResearchMCP
Enhancing research capabilities within Claude Desktop
Providing web search functionality to AI models
Automating research tasks
Building custom research tools
FAQ from ResearchMCP
What is the Model Context Protocol (MCP)?
What is the Model Context Protocol (MCP)?
MCP is a protocol that allows applications to interact with AI models by providing contextual information.
What search engines are supported?
What search engines are supported?
Currently, ResearchMCP supports Brave Search and Tavily Search.
What are the system requirements?
What are the system requirements?
You need Deno installed for local development and Docker for containerized deployment.
How do I set up the API keys?
How do I set up the API keys?
You need to set the BRAVE_API_KEY and optionally the TAVILY_API_KEY environment variables.
What languages are supported?
What languages are supported?
Due to limitations with the Brave Search API, English queries are recommended. Non-Latin characters may cause encoding errors.