Overview
What is Deep_research?
Deep_research is an agent-based tool that provides web search and advanced research capabilities. It leverages HuggingFace’s smolagents and is implemented as an MCP (Model Context Protocol) server, designed for users who need automated information gathering from multiple sources.
How to use Deep_research?
After cloning the repository, create a virtual environment with uv, install dependencies with uv sync, and set the required environment variables (OPENAI_API_KEY, HF_TOKEN, SERPER_API_KEY) in a .env file. Start the MCP server by running uv run deep_research.py.
Key features of Deep_research
- Web search and information gathering
- PDF and document analysis
- Image analysis and description
- YouTube transcript retrieval
- Archive site search
Use cases of Deep_research
- Automating web research for reports or articles
- Extracting and summarizing content from PDF documents
- Analyzing images and generating descriptions
- Fetching and processing YouTube video transcripts
- Searching archived web pages for historical data
FAQ from Deep_research
What are the dependencies and runtime requirements?
Python 3.11 or higher and the uv package manager are required. The server also needs API keys for OpenAI, HuggingFace, and SerpAPI.
How do I obtain the required API keys?
An OpenAI API key, a HuggingFace token, and a SerpAPI key (sign up at Serper.dev) must be set as environment variables in a .env file.
How do I start the server?
Run uv run deep_research.py from the project root after installing dependencies and configuring environment variables.
What tools or components are included?
Key scripts include text_web_browser.py, text_inspector_tool.py, visual_qa.py for image analysis, and mdconvert.py for converting files to Markdown.
