assafelovic/GPT Researcher
An autonomous agent that conducts deep research on any data using any LLM providers
GPT Researcher is an open-source autonomous agent that runs deep research on any topic and writes a cited report, with any LLM and search provider. It plans sub-questions, scrapes and ranks sources, and assembles the findings instead of leaving you with dozens of open tabs.
Our Review
GPT Researcher, created by Assaf Elovic, was one of the first open-source deep-research agents and has grown to 27,000 GitHub stars. You give it a question, it runs the full search-scrape-synthesize loop autonomously, and hands back a cited report, the same workflow paid deep-research features now charge a subscription for.
What GPT Researcher does:
- •Autonomous research loop plans sub-questions, runs web searches, scrapes pages, and filters for relevance without step-by-step prompting.
- •Cited reports outputs a structured report with inline sources, so you can verify every claim.
- •Bring any model and search works with OpenAI, Anthropic, Google, local models, and search backends like Tavily, Bing, or SearXNG.
- •Research your own data point it at local documents or a vector store, not only the open web.
- •Deep and detailed modes choose quick summaries or longer multi-agent deep-research runs for thorough reports.
- •MCP server and frontend run it as an MCP server for other agents, or use the built-in web UI.
Getting started:
Install with pip install gpt-researcher, set your LLM and search API keys, and call the researcher with a query; or run the web UI from the repo. Docs at docs.gptr.dev.
Limitations:
GPT Researcher's quality and cost depend on the models and search APIs you plug in: deep runs make many LLM and search calls, so a thorough report has real token and API cost. It is a developer tool, so you wire in keys and run it yourself rather than using a polished consumer product. Results still need human review, since web sources vary in reliability and the agent can surface low-quality pages.
Our Verdict
GPT Researcher is a proven open-source way to automate deep research in 2026. If you regularly need a sourced briefing on a topic and do not want to pay per seat for a closed deep-research product, this agent runs the whole search-and-synthesize loop and gives you a cited report, with 27,000 stars and an Apache-2.0 license.
For developers, it installs as a Python package and exposes the researcher programmatically, so you can embed automated research in your own apps, pipelines, or agents. You control the models, the search backends, and the cost, and the MCP server lets other agents call it as a tool.
Skip GPT Researcher if you want a polished, zero-setup consumer experience; a hosted tool like Perplexity is friendlier for ad-hoc questions. If your research is shallow or one-off, a single model call with web search may be cheaper than a full deep-research run.
Frequently Asked Questions
What is GPT Researcher?
GPT Researcher is an open-source autonomous agent, created by Assaf Elovic, that conducts deep research on a topic and produces a cited report. Given a question, it plans sub-queries, searches the web or your own data, scrapes and filters sources, and synthesizes the findings, all using whichever LLM and search providers you configure.
Is GPT Researcher free and open source?
Yes. GPT Researcher is released under the Apache-2.0 license and is free and open source as of 2026. You install it from PyPI as gpt-researcher at no licensing cost. The expenses you take on are the LLM provider and search API calls it makes, which grow with how deep and broad your research runs are.
How does GPT Researcher work?
GPT Researcher runs an autonomous loop. It breaks your question into sub-questions, issues searches, scrapes and ranks the resulting pages, and filters for relevant, credible content. It then synthesizes the gathered material into a structured report with citations. Deeper modes use multiple agents and more passes for longer, more thorough results.
How is GPT Researcher different from Perplexity?
Perplexity is a polished hosted answer engine you use through a website or app. GPT Researcher is an open-source agent you run yourself, with full control over models, search backends, prompts, and cost, and it can research your private data. Choose Perplexity for quick, no-setup answers; choose GPT Researcher to embed automated research in your own stack.
Can GPT Researcher use my own documents?
Yes. As of 2026, GPT Researcher can run research over your local documents or a connected vector store, not only the open web. That lets you generate cited reports grounded in internal files, and you can mix local and web sources in a single run. You configure the data sources alongside your model and search keys.
How do I install GPT Researcher?
Visit the GitHub repository at https://github.com/assafelovic/gpt-researcher for installation instructions.
What license does GPT Researcher use?
GPT Researcher uses the Apache-2.0 license.
What are alternatives to GPT Researcher?
Explore related tools and alternatives on My AI Guide.
Open source & community-verified
Apache-2.0 licensed: free to use in any project, no strings attached. 27,411 developers have starred this, meaning the community has reviewed and trusted it.
Reviewed by My AI Guide for relevance, quality, and active maintenance before listing.
Topics