deep-research
An AI-powered research assistant that performs iterative, deep research on any topic by combining search engines, web scraping, and large language models. The goal of this repo is to provide the simplest implementation of a deep research agent — e.g. an agent that can refine its research direction overtime and deep dive into a topic.
- stars
- 19,684
- 30-day movement
- +124/day
- Related entries
- 60
- Connections
- 1
deep-research is a TypeScript AI research agent that combines search engines, web scraping, and large language models to research a topic iteratively. It is built as a deliberately simple implementation of a deep research agent that refines its research direction over time and dives deeper into a topic.
You want a minimal, readable codebase showing how a deep research loop (search, scrape, refine, repeat) is wired together, rather than a heavyweight framework.
Use it to
- Run iterative deep research on a topic
- Study a minimal deep research agent implementation
- Extend or fork the research loop in TypeScript
- Experiment with LLM-driven search and scraping pipelines
For Developers exploring or building LLM research agents
- Role
- agent-app
- Language
- TypeScript
- Licence
- MIT
- Forks
- 2,000
- Open issues
- 72
- Last push
- 2026-04-11