# gpt-researcher

> gpt-researcher is an open-source agent for research. Open-source research agent that writes cited reports from the web and your documents. It is available on Web, API, CLI and self-hosted.

- **Kind:** Agent
- **Pricing:** Open source
- **Platforms:** Web, API, CLI, self-hosted
- **Website:** https://gptr.dev
- **Repository:** https://github.com/assafelovic/gpt-researcher
- **GitHub stars:** 30,746
- **License:** Apache-2.0
- **Categories:** research, web search and scraping
- **Listing updated:** 2026-11-14

## About

A planner turns your question into a set of research questions and runs a researcher for each one in parallel. Context filtering with Jev, BM25 or embeddings keeps only the passages that answer the question before the writer sees them, and in deep-research mode findings seed new questions, so a run explores a tree of subtopics.

More than 20 retrievers are available, including Tavily, Google, Bing, Exa, arXiv and PubMed, and these can be combined with MCP servers or your own documents. Every claim in the report traces back to a source, and reports export as markdown, PDF, docx and more. An example shows a report built from 23 sources in 47 seconds.

It is Python, Apache-2.0, and a separate MCP server exposes tools such as deep_research and quick_search to Claude, Cursor and n8n. The site cites a first-place result on a Carnegie Mellon benchmark.

## Features

- **Parallel researchers:** A planner assigns a researcher to each sub-question.
- **20+ retrievers:** Web, academic and custom sources, MCP servers and your documents.
- **Context filtering:** Keeps only relevant passages before they reach the model.
- **Deep research mode:** Findings generate new questions across subtopics.
- **Cited reports:** Claims link back to sources.
- **MCP server:** Exposes research tools to Claude, Cursor and n8n.

## Good for

- Draft a sourced report on a market or topic
- Search academic papers on arXiv and PubMed
- Research your own documents alongside the web
- Call research from another agent through MCP

## FAQ

### What licence is it?

Apache-2.0.

### What sources can it search?

Over 20 retrievers such as Tavily, Google, Bing, Exa, arXiv and PubMed, plus MCP servers and local documents.

### What formats can reports be exported in?

Markdown, PDF, docx and others.

### Can other agents call it?

Yes, through the gptr-mcp server.

### Is gpt-researcher free?

Yes. gpt-researcher is open source, so you can use and self-host it for free.

### What are the best alternatives to gpt-researcher?

The closest alternatives to gpt-researcher are deer-flow, DeepCode, ai-agent-book, Vibe-Trading and BettaFish.

## Alternatives

- [deer-flow](https://ai-powered.com/agents/deer-flow): deer-flow is an open-source agent for ai models and llm tooling.
- [DeepCode](https://ai-powered.com/agents/deepcode): DeepCode is an open-source agent for coding.
- [ai-agent-book](https://ai-powered.com/agents/ai-agent-book): ai-agent-book is an open-source hosted agent for knowledge and memory.
- [Vibe-Trading](https://ai-powered.com/agents/vibe-trading): Vibe-Trading is an open-source agent for finance and payments.
- [BettaFish](https://ai-powered.com/agents/bettafish): BettaFish is an open-source agent for analytics and data.
- [awesome-ai-agents](https://ai-powered.com/agents/awesome-ai-agents): awesome-ai-agents is a free hosted agent for research.
- [agents-towards-production](https://ai-powered.com/agents/agents-towards-production): agents-towards-production is a free open-source agent for ai models and llm tooling.
- [Auto-claude-code-research-in-sleep](https://ai-powered.com/agents/auto-claude-code-research-in-sleep): Auto-claude-code-research-in-sleep is an open-source agent for research.

---
Source: https://ai-powered.com/agents/gpt-researcher
