GPT Researcher Review 2026: Open-Source AI Research Agent Guide
VerifiedA review of GPT Researcher, a free open-source agent that builds cited research reports from the web and local documents.
Categories & Tags
GPT Researcher is an open-source AI research agent that plans research questions, gathers sources, and publishes a cited report from the web and local documents using the LLM you choose. As an AI research assistant tool and AI deep research agent, it is free and open source under Apache-2.0, with code at github.com/assafelovic/gpt-researcher. If you are comparing options, see our list of AI research tools.
Quick Navigation
Quick Summary
Overall Rating 4.3/5 Best For It suits developers and technical researchers who want to run their own research agent with control over LLM and search providers. It is less suited to non-technical users who prefer a hosted app. Pricing Yes, the software is free and open-source under Apache-2.0. You pay only for the LLM and search API usage you choose. Free Plan Yes Ease of Use 3.5/5
What GPT Researcher Is
GPT Researcher is an autonomous agent that conducts deep research on any data using any LLM provider. It uses a planner agent to generate research questions, execution agents to gather relevant information, and a publisher that aggregates findings into a comprehensive report. Unlike Elicit, which is a hosted research assistant, the tool is open-source code you install and run yourself. The code is Apache-2.0 licensed and available on GitHub at assafelovic/gpt-researcher, with docs at docs.gptr.dev.
Who the free AI research assistant suits
This AI agent research assistant suits people who want a cited report assembled from web and local documents with a chosen LLM. It is not recommended for academic or research papers, per its disclaimer that output is not academic advice.
- Agent developers Developers building research agents can read and modify the planner, execution, and publisher code. The project is open source under Apache-2.0.
- Analysts and content teams Teams that need detailed reports with citations can generate reports that exceed 2,000 words and aggregate over 20 sources. Reports export to PDF, Word and other formats.
- Private document research Teams that want to research private or local documents can point GPT Researcher at local files as well as the web. It also supports JavaScript-enabled web scraping.
- Provider choice People who want to choose their own LLM and search provider can swap models and retrievers. The default setup needs an OpenAI API key and a Tavily API key, but other providers and local models work via a custom base URL.
MCP Support and Key Features
Planner and execution agents
A planner agent generates research questions, and execution agents gather relevant information for each one. A publisher then aggregates all findings into a comprehensive report. This is narrower than a general-purpose agent framework like CrewAI, because GPT Researcher is built only for research.
Long cited reports
Reports can exceed 2,000 words and aggregate over 20 sources. Output is detailed and factual, with citations. Reports export to PDF, Word and other formats.
Web and local documents
It researches the web and local documents. It includes smart image scraping and filtering, AI-generated inline images, and JavaScript-enabled web scraping. Smart context filtering keeps only useful context, with no extra API key or embeddings required by default.
Deep Research mode
Deep Research mode runs recursive research in breadth and depth with concurrent processing. It takes about 5 minutes per deep research and costs about $0.40 per research using o3-mini on high reasoning effort. The software itself is free; you pay only for the LLM and search API usage you choose.
GPT Researcher MCP support
GPT Researcher MCP support lets it act as an MCP client to connect specialized data sources such as GitHub repositories, databases and custom APIs. This extends research beyond web search and local files. It is part of the open-source project on GitHub.
Choose LLM and retriever
You can use OpenAI or other LLM providers and OpenAI-compatible APIs, including local models via a custom base URL. Tavily is the default search retriever, and the project also supports Google, SearchApi, SerpApi, Serper, Searx, DuckDuckGo, Semantic Scholar, PubMed Central, GroundRoute, fastCRW, and custom retrievers. Tavily has a free plan with 1,000 API credits per month.
Is GPT Researcher Free to Use?
Yes, GPT Researcher is free. It is open source under Apache-2.0, so you can run deep research AI free of licence fees. You pay only for LLM and search API usage; Tavily, the default search, has a free plan with 1,000 credits/month, and a Deep Research run costs about $0.40 with o3-mini on high reasoning effort. Your costs depend entirely on your chosen LLM and search API usage.
| Cost Item | Price | Notes |
|---|---|---|
| GPT Researcher software | Free | Apache-2.0 open source |
| LLM API (e.g. OpenAI) | Pay per use | You pay your provider directly |
| Search API (Tavily default) | Free plan: 1,000 credits/month | Paid plans available for more usage |
| Deep Research run | About $0.40 per run | Using o3-mini on high reasoning effort |
Strengths and Limitations
- Cited research reportsProduces detailed, factual reports with citations that can exceed 2,000 words and aggregate over 20 sources.
- Web and local researchResearches the web and local documents, with smart image scraping and filtering plus JavaScript-enabled scraping.
- Multi-agent planningPlanner and execution agents generate research questions, gather information, and a publisher aggregates findings into a report.
- Flexible providersWorks with many LLM providers and search retrievers, including local models via a custom base URL. Smart context filtering needs no extra API key or embeddings by default.
- Setup requirementsNeeds Python 3.10+ and API keys for an LLM and a search provider. This makes initial setup less simple than a hosted tool.
- Usage-based costsThe software is free, but your costs depend on the LLM and search API usage you choose. Deep Research runs add up at about $0.40 each with o3-mini high.
- Self-Hosted Setup RequiredYou install it via pip or run it from the repo with a local server, and you control the deployment yourself. This means more setup than a hosted tool.
- Academic disclaimerThe project's own disclaimer says its output is not academic advice and is not a recommendation for use in academic or research papers.
Use Cases Including AI for Summarizing Research Papers
Market and competitor reports
Generates detailed reports with citations on markets and competitors. For broader options, see AI tools for market research.
AI to Search for Research Papers
The AI research papers finder can search Semantic Scholar and PubMed Central as retrievers. It gathers sources and writes a cited report that summarizes what it found. Its own disclaimer says output is not academic advice and not a recommendation for use in academic or research papers.
Local document research
Researches local documents as well as the web. Smart context filtering keeps only useful context without extra API keys or embeddings by default.
Build research into your app
Use the pip package or MCP support to connect specialized data sources like GitHub repositories, databases, and custom APIs. This lets you add research features to your own application.
Installing and Creating Your First Report
Make sure you have Python 3.10 or newer installed.
Install it with pip install gpt-researcher, as with many AI research tools GitHub hosts as Python packages.
Set your OpenAI and Tavily API keys, or configure another LLM provider and retriever.
Run a research query from the package or the local server.
Export the report to PDF, Word, or another supported format.
Is It Worth the Cost?
GPT Researcher is worth using if you need cited, multi-source research reports and are comfortable running Python with API keys. It is free and open source, and it works with many LLM providers and search retrievers, though your costs depend on usage. For a broader comparison, see our list of open-source AI tools.
Comparison with AI Deep Research Tools
GPT Researcher is an open-source AI tool for deep research that runs with your chosen LLM, while Perplexity Deep Research, Elicit and NotebookLM are hosted apps. Choosing a deep research AI tool therefore comes down to whether you prefer a self-hosted setup with usage-based costs or a hosted service.
GPT Researcher vs Perplexity Deep Research
Perplexity Deep Research is a hosted product you use through a browser. GPT Researcher is open-source code you run with your own LLM and search API keys. Both produce cited research reports, but only it lets you inspect and modify the underlying agent code.
GPT Researcher vs Elicit
Elicit is a hosted research assistant focused on finding and summarizing academic papers. GPT Researcher is a self-run open-source agent that researches the web and local documents with the LLM and search providers you choose. Elicit is simpler to start; it gives you more control over sources and models.
GPT Researcher vs NotebookLM
NotebookLM is a hosted tool for working with documents you upload. GPT Researcher is open-source software you install and run, and it can search the live web as well as local files. NotebookLM is easier for non-technical users; it requires Python and API keys.
Frequently Asked Questions
Is GPT Researcher free?
Yes, the software is an AI research assistant free to download and use, and it is open-source under the Apache-2.0 license. You pay only for the LLM and search API usage you choose, and the default Tavily search API has a free plan with 1,000 credits per month.
What is GPT Researcher MCP support?
It works as an MCP client to connect specialized data sources such as GitHub repositories, databases and custom APIs. This lets the agent pull information from systems beyond standard web search.
How do I install GPT Researcher from GitHub?
Install GPT Researcher with pip install gpt-researcher on Python 3.10+, or clone the AI research assistant GitHub repository at assafelovic/gpt-researcher. The code is at github.com/assafelovic/gpt-researcher under Apache-2.0.
Can GPT Researcher search academic research papers?
Yes, it supports search retrievers including Semantic Scholar and PubMed Central for academic sources. The project's own disclaimer says its output is not academic advice and is not a recommendation for use in academic or research papers.
What is a good GPT Researcher alternative?
Hosted options such as Perplexity Deep Research, Elicit and Consensus are good alternatives if you prefer not to run code yourself. They are hosted apps, while the agent is self-run open-source software with your own LLM and search choice.
Key Takeaways
- GPT Researcher is a free, open-source autonomous agent that produces cited research reports, often over 2,000 words from more than 20 sources.
- It researches both the web and local documents, with JavaScript-enabled scraping and smart context filtering that needs no extra API key by default.
- Deep Research mode runs recursive, concurrent research in about 5 minutes and costs about $0.40 per research using o3-mini on high reasoning effort.
- It fits alongside other AI agents as a self-run option where you control the LLM and search providers.
GPT Researcher Alternatives
- For researchers who need traceable answers, Consensus is a hosted platform that answers questions with findings and citations from scientific papers.
- If you want to see whether a citation supports or contrasts a paper's claims, Scite is a hosted platform built for that citation-level check.
- When your work requires deep academic literature searches, Undermind is a hosted platform designed for that depth.
- Developers who prefer an API over a browser interface can use Exa, an AI-native web search engine offered as an API.
For more options, see our lists of AI tools for researchers and Perplexity alternatives.
Bottom Line: GPT Researcher is a strong pick if you want an open-source research agent you run yourself with your own LLM and search keys. It produces long cited reports and supports web, local documents and MCP data sources. The trade-off is setup: you need Python and API keys, which is why it rates 3.5/5 for ease of use.
Last Reviewed: November 2026
Key Features
Planner And Execution Agents
Long Cited Reports
Web And Local Documents
Deep Research Mode
Gpt Researcher Mcp Support
Choose Llm And Retriever
Use Cases
Market and competitor reports
AI to Search for Research Papers
Local document research
Build research into your app
Pros & Cons
Pros
- Cited research reports
- Web and local research
- Multi-agent planning
- Flexible providers
Cons
- Setup requirements
- Usage-based costs
- Self-Hosted Setup Required
- Academic disclaimer
Quick Answers
Is GPT Researcher free?
Yes, the software is free and open-source under Apache-2.0. You pay only for the LLM and search API usage you choose.
Who is GPT Researcher best for, especially developers or teams that want self-hosting?
It suits developers and technical researchers who want to run their own research agent with control over LLM and search providers. It is less suited to non-technical users who prefer a hosted app.
How easy is GPT Researcher to set up and use for someone comfortable with Python?
It has a moderate learning curve because it needs Python and API keys. We rate ease of use 3.5/5.
GPT Researcher
AI Research Tools
Pricing Plans
Free
Basic features included
GPT Researcher software
FreeApache-2.0 open source
LLM API (e.g. OpenAI)
Pay per useYou pay your provider directly
Search API (Tavily default)
Free plan: 1,000 credits/monthPaid plans available for more usage
Deep Research run
About $0.40 per runUsing o3-mini on high reasoning effort