I like building things that think. These days I'm working on Ora, GPT Researcher, Tavily and a few side projects.
You can find me at assafe.com, on X and on LinkedIn.
|
An autonomous agent that conducts deep research on any data using any LLM providers 
|
Score any site's agent readiness from your terminal or CI. The open source command line tool for Ora. 
|
|
An MCP server that gives Claude, Cursor and other MCP clients deep research, quick search and report writing tools. 
|
The official Python SDK for Tavily's search, extract, crawl, map and research APIs. 
|
|
Helps agents like OpenClaw find and learn new skills on their own. 
|
A memory layer that helps AI agents give personal, context aware answers. 
|
- Ora checks how well AI agents can find, read and use a website. Its open source tools are at github.com/ora, including ax, which scores agent readiness from your terminal or CI.
- GPT Researcher is the open source deep research agent I created back in 2023, and it still has an active community to this day. The MCP server lets Claude, Cursor and other agents use it.
- Tavily is the search engine for AI agents that I co-founded. The Python SDK is the easiest way to start using it.
- Side projects include skyll, which helps agents learn new skills, and tovana, a memory layer for agents.
I started GPT Researcher in 2023 as the first open source deep research agent. You give it a question, it plans the research, reads dozens of sources in parallel and writes a report with citations. The community has kept improving it every week since.
To try it, start with the quickstart or install it from PyPI. You can also connect it to your own agent with the MCP server. Everything else is on gptr.dev and docs.gptr.dev, and you're welcome to say hi on Discord.
Carnegie Mellon's DeepResearchGym (May 2025) tested deep research systems on 1,000 complex questions. GPT Researcher came first, ahead of Perplexity, OpenAI, OpenDeepSearch and Hugging Face.
| Metric | GPT Researcher |
|---|---|
| Citation precision | 85.36% |
| Citation recall | 90.82% |
| Report clarity | 83.70% |
| Report insightfulness | 78.01% |
| Key point recall | 64.67% |
It also shows up in these papers:
- Deep Research Comparator: A Platform For Fine-grained Human Annotations of Deep Research Agents, Chandrahasan et al., July 2025
- Deep Researcher with Test-Time Diffusion, Han et al., Google, July 2025
- HLTCOE at LiveRAG: GPT-Researcher using ColBERT retrieval, Duh et al., Johns Hopkins HLTCOE, June 2025
- Chain of Ideas: Revolutionizing Research in Novel Idea Development with LLM Agents, Li et al., October 2024
- Providing domain knowledge for process mining with ReWOO-based agents, Vogt, van der Putten and Reijers, GenAI4PM at ICPM 2024
- BioKGBench: A Knowledge Graph Checking Benchmark of AI Agent for Biomedical Science, Lin et al., July 2024
- LLM4DESIGN: An Automated Multi-Modal System for Architectural and Environmental Design, Chen et al., July 2024
- Differential Privacy of Cross-Attention with Provable Guarantee, Gu et al., July 2024
- Exploring Large Language Model based Intelligent Agents: Definitions, Methods, and Prospects, Cheng et al., January 2024