DeerFlow

DeerFlow is an in-depth research framework that aims to combine language models with dedicated tools such as web search, crawlers and Python execution to promote in-depth research. The project originated from the open source community, emphasizes contribution and feedback, has a variety of flexible functions, suitable for various research needs.

Demand population:

"This product is suitable for scientific researchers, developers and students because it combines advanced research tools and models, which can effectively improve research efficiency and information acquisition capabilities, and meet a variety of academic and practical needs."

Example of usage scenarios:

Use DeerFlow to generate a research report on traditional Nanjing food.

Transform the generated research reports into podcast audio through the TTS feature.

Use DeerFlow for code analysis and execution to improve software development efficiency.

Product Features:

Supports integration with multiple LLM models, adapted to open source models and OpenAI API interfaces.

Provides a variety of search engine support, such as Tavily, DuckDuckGo and Brave Search, for easy information retrieval.

Integrate text-to-speech functionality to generate voice using high-quality TTS API.

Supports human-computer collaboration and allows users to modify research plans in natural language.

Provides post-reporting functions, supporting block editing and AI-assisted polishing.

Automatically generate podcast audio and simple PowerPoint presentations.

The modular multi-agent system architecture optimizes the research and code analysis process.

Equipped with Web UI and console UI to enhance user interaction experience.

Tutorials for use:

Cloning project code: git clone https://github.com/bytedance/deer-flow.git

Enter the project directory: cd deer-flow

Install dependencies: uv sync

Configure the .env file and add the API key.

Run the main program: uv run main.py