GPT Researcher
GPT Researcher is an open-source deep research agent that investigates a topic across the web and your local documents and writes a detailed, cited report. A planner agent generates research questions, execution agents gather and summarise sources for each question, and a publisher aggregates the findings into a final report, typically drawing on more than 20 sources and often exceeding 2,000 words.
It suits analysts, researchers and product teams who want a first draft of a literature or market review, a due-diligence summary or a technical background report, and developers who want to embed research capabilities in their own apps. It supports custom and local models through OpenAI-compatible endpoints, can combine web search with MCP data sources such as GitHub or databases, exports to PDF and Word, and comes with both a lightweight and a Next.js front end. It can also be installed as a Claude skill.
Getting started: install with `pip install gpt-researcher` to use it as a library, or clone the repository (Python 3.12+), set `OPENAI_API_KEY` and a search provider key such as `TAVILY_API_KEY`, and run the server with uvicorn. `OPENAI_BASE_URL` points it at other OpenAI-compatible providers.
Features
Why choose it
Trade-offs
Maintenance check
- Last commit
- 2026-11-09
- Commits, last 90 days
- 100+
- Releases, last 90 days
- 3
- Latest release
- v3.8.0 (2026-11-09)
- Open issues
- 17
- Recent issues with at least one comment
- 3/5
Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.
Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.
Most discussed open issues
- #1764 Research is getting stucked every other time while giving no meaningful logs / ui hints indicating what it was attempting to do6 comments · 2026-05-07
- #1582 GPL License Alternative PDF Scraper Implementation2 comments · 2026-01-05
- #2277 PyMuPDFScraper returns empty titles for PDFs without document metadata1 comments · 2026-11-09
Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.
Who it fits
- You want an Apache-2.0 agent that plans sub-questions, searches the web and local documents in parallel, and writes 2,000+ word reports with citations.
- You want to use it as a Python library, standalone web app, or MCP-connected research skill with PDF/Word export.
- You want a general-purpose super-agent harness with a bash sandbox, persistent memory, and Feishu/WeCom/Slack bots; DeerFlow is built for that.
- You do not want per-report search API (e.g. Tavily) and multi-call LLM token costs.
How it differs
GPT Researcher focuses specifically on parallel web/document research and cited report generation (Apache-2.0). DeerFlow 2.0 is a broader LangGraph-based agent harness with a Docker sandbox, filesystem, sub-agents, and chat-app connectors (MIT).
GPT Researcher is a model-agnostic orchestration pipeline that works with OpenAI, Anthropic, or local models. MiroThinker provides open-weights research-agent models (30B/235B), the MiroFlow framework, and training datasets.
Compatibility
Quick start
Use cases
Alternatives
Featured in these lists
Comments
No comments yet. Be the first!