RUC-NLPIR/WebThinker
WebThinker is a deep-research agent framework that enables large reasoning models to conduct multi-step web exploration and synthesize findings into structured reports.
Not currently ranked — collecting fresh signals.
star history
WebThinker extends large reasoning models such as DeepSeek-R1 and QwQ with deep research capabilities by orchestrating web search, page retrieval, and information synthesis. It generates Reflect tokens to guide exploration and produces comprehensive research reports from web sources. The project provides fine-tuned model checkpoints (7B to 32B) trained on research trajectories and supports integration with search APIs for real-time information gathering.
Frequently asked
- What is RUC-NLPIR/WebThinker?
- WebThinker is a deep-research agent framework that enables large reasoning models to conduct multi-step web exploration and synthesize findings into structured reports.
- Is WebThinker open source?
- Yes — RUC-NLPIR/WebThinker is open source, released under the MIT license.
- What language is WebThinker written in?
- RUC-NLPIR/WebThinker is primarily written in Python.
- How popular is WebThinker?
- RUC-NLPIR/WebThinker has 1.5k stars on GitHub.
- Where can I find WebThinker?
- RUC-NLPIR/WebThinker is on GitHub at https://github.com/RUC-NLPIR/WebThinker.