gptr-mcp

MCP.Pizza Chef: assafelovic

Built on GPT Researcher, it takes a question, hunts down and cross-checks many sources, then hands back a written report along with the sources it leaned on, rather than a raw list of links. Expect roughly thirty seconds for a deep search. It runs on your own machine and needs paid keys from OpenAI and Tavily. Worth knowing before you invest time: the project has had no new code since November 2025, so treat it as stable but unattended.

Web/Research

Writing

Use This MCP server To

Research an unfamiliar topic and get a written summary back Compare three products and see where their claims disagree Gather background reading before writing a report Check which sources actually back up a claim Run a quick web search when a full report is overkill

README

🔍 GPT Researcher MCP Server

Why GPT Researcher MCP?

While LLM apps can access web search tools with MCP, GPT Researcher MCP delivers deep research results. Standard search tools return raw results requiring manual filtering, often containing irrelevant sources and wasting context window space.

GPT Researcher autonomously explores and validates numerous sources, focusing only on relevant, trusted and up-to-date information. Though slightly slower than standard search (~30 seconds wait), it delivers:

✨ Higher quality information
📊 Optimized context usage
🔎 Comprehensive results
🧠 Better reasoning for LLMs

💻 Claude Desktop Demo

gptr-mcp FAQ

Is this project still maintained?

No new code has landed since November 2025. It still works as written, but nobody is actively updating it.

Which apps does this work in?

Claude Desktop is the documented setup, and it can also run over a web connection for tools like n8n.

Do I need to pay for anything?

Yes. You supply your own OpenAI and Tavily keys, and both services charge for use.

Can I use this to research a market I know nothing about?

Yes, that is its main job. It reads and filters sources first, then writes the report.

How long does a deep search take?

About thirty seconds. Slower than a plain web search, but it returns far less irrelevant material.