Web search API built for AI agents.
GPT Researcher
Open-source agent that researches the web and writes reports.
SharkScore
#8 in Research & search
Why it scores 67.7
GPT Researcher has a SharkScore of 67.7/100, ranked #8 of 42 in Research & search (updated 2026-11-19).
Each part compares GPT Researcher with other listings in Research & search. Evidence covers 87% of the score's weight.
- Adoption
- 46 · 30%
- User reviews
- no data · 25%
- Maintenance
- 100 · 20%
- Trust & transparency
- 80 · 15%
- Momentum
- no data · 10%
Paying never changes this score. How ranking works · method v2026-11-18.2
RankShark review of GPT Researcher
Automated evaluation by RankShark · 18 Nov 2026 · method v1.0 · How we review
We were not paid for this review.
As of 2026-11-18, GPT Researcher is a self-hosted, open-source web and document research agent; there is no hosted SaaS subscription, and provider costs are separate. [E1, E2]
GPT Researcher plans research questions, searches web sources or customer documents, filters passages and writes reports with citations. It can be run as a Python package or self-hosted server. [E1, E2, E3]
Best for
- Developers who want to run a research agent themselves and choose their LLM and retriever. · home page, pricing page
- Teams that want reports based on web sources, their own documents, or both. · home page, docs
- People who want to connect research tools through MCP clients. · home page
Not for
- Teams looking for a hosted SaaS offering from GPT Researcher; the pricing page says it does not currently provide one. · pricing page
Key features
- Parallel research. The vendor says a planner splits a question into sub-questions and researchers work on them in parallel. · home page, docs
- Web and document sources. The vendor says it can research web sources and local documents, including hybrid research that uses both. · home page, pricing page, docs
- Cited reports. The vendor says reports include citations tracing claims to sources and can be produced in Markdown, PDF or Word. · home page, docs
- Deep research. The vendor says findings can seed new questions, exploring a topic through branching subtopics. · home page, docs
- Context filtering. The vendor says Jev filters sources for relevant passages and can fall back to local keyword ranking. · home page, docs
- MCP sources. The vendor says MCP can connect sources such as GitHub, databases and user-provided APIs. · docs
Pros
- The pricing page states that the open-source package is free and that GPT Researcher itself imposes no usage limits. · pricing page
- It supports research over local documents as well as web sources, including a combined mode. · home page, pricing page, docs
- The vendor says users can choose supported LLMs and retrievers, including local models and free search options. · home page, pricing page
- The vendor says the official MCP server works with clients including Claude Desktop, Cursor, ChatGPT custom GPTs and n8n. · home page
Cons
- There is no hosted SaaS from the vendor, and users pay their LLM and search providers separately. · pricing page
- The homepage and pricing page state an MIT license, while the documentation states Apache 2.0; the pages do not agree on the license. · home page, pricing page, docs
- Whether customer data is used for model training is not stated on the pages we checked.
Pricing
| Plan | Price | Limits |
|---|---|---|
| Open Source | $0 | Full GPT Researcher Python package on PyPI; official gptr-mcp server; self-hosted; use supported LLMs and retrievers. No usage limits imposed by GPT Researcher; provider limits may apply. |
Checked 18 Nov 2026 · source
- Free plan
- Open-source software costs $0. There is no SaaS subscription or rate limit imposed by GPT Researcher; users pay their chosen LLM and search providers directly. · pricing page
- Free trial
- Not stated (checked 18 Nov 2026)
- Trains on customer data
- Not stated (checked 18 Nov 2026)
- Data retention
- Not stated (checked 18 Nov 2026)
- Data regions
- Not stated (checked 18 Nov 2026)
- Data processing agreement
- Not stated (checked 18 Nov 2026)
- AI models used
- The vendor says it supports any LLM, naming OpenAI, Anthropic, Google, Mistral, DeepSeek and local models via Ollama. · home page
- API
- The vendor describes a FastAPI server and says its endpoints are documented at gptr.dev/openapi.json. · home page
- SDKs
- Not stated (checked 18 Nov 2026)
- Webhooks
- Not stated (checked 18 Nov 2026)
- Platforms
- The vendor says it can be run on a user's machine, server or container. Its MCP server is listed for Claude Desktop, Cursor, ChatGPT custom GPTs, n8n and other MCP-capable clients. · home page, pricing page
- Security claims
- No security certification stated (checked 18 Nov 2026)
What we measured
- GitHub stars · now
- 30,760
- Latest release · now
- v3.8.0 on 2026-11-09
- Package downloads · 30 days
- 33,093
- Tranco traffic rank · list of the day
- #1,604,649
How we tested · sources
- Hourly uptime and response-time checks of the product's URL.
- Repository, release and download data from GitHub, npm, PyPI and Hugging Face where they exist.
- We did not sign up, pay, or use the product as a customer.
- home page: https://gptr.dev/ · fetched 18 Nov 2026
- pricing page: https://gptr.dev/pricing · fetched 18 Nov 2026
- docs: https://docs.gptr.dev/ · fetched 18 Nov 2026
- privacy page: https://gptr.dev/policy · fetched 18 Nov 2026
- 18 Nov 2026: Plans or prices changed on the pricing page.
Something wrong? Report an error with a public link that shows it; corrections re-run the review.
Top alternatives to GPT Researcher
See all- 69.7ToolFreemium
- 68.9
Web search and extraction API for AI agents.
ToolFreemium · from $30/mo - 68.5
Answers with sources, straight from the web.
ToolFreemium - 66.3
Ad-free paid search engine with an AI assistant.
ToolPaid · from $4.5/mo