DeerFlow 2.0 - ByteDance's open-source super intelligent agent framework

What is DeerFlow 2.0?

DeerFlow 2.0 is an open-source super agent framework from ByteDance. It adopts a "main agent + 11-layer middleware chain + dynamic sub-agents" architecture, achieving multi-agent collaboration through LangGraph. The framework includes over 10 pluggable skills such as deep research, data analysis, and audio/video creation, and supports Docker/Kubernetes isolated sandbox execution. It natively adapts to Lark, Telegram, and Slack, and can run without a public IP address. Version 2.0 upgrades from a single research tool to a general-purpose agent runtime, supporting complex task processing from minutes to hours, and can generate websites, PPTs, comics, and other content with a single click.

Main features of DeerFlow 2.0

• Intelligent agent architectureIt adopts a collaborative mode of overall planning by a main intelligent agent and parallel execution by dynamic sub-intelligent agents to achieve efficient decomposition and collaborative processing of complex tasks.
• Pluggable skill systemIt includes more than 10 skill modules such as in-depth research, data analysis, PPT generation, and web design, and supports users to customize and extend new capabilities through the MCP interface and Python functions.
• Sandbox ImplementationIt offers three running modes: local, Docker, and Kubernetes. Based on ByteDance AIO Sandbox, it creates an independent container environment for each task to ensure execution security and resource isolation.
• Context EngineeringThe system systematically addresses the issue of insufficient context windows in long-running tasks through mechanisms such as automatic summary compression, external storage of intermediate results, and subtask rate limiting.
• Long-term memoryIt supports persistent storage of user profiles, preferences, and accumulated knowledge across sessions, and all data is stored locally and is completely under the user's control.
• IM channel integrationIt is natively compatible with the three major instant messaging platforms, Lark, Telegram, and Slack, and can interact with intelligent agents via commands without the need for a public IP address.
• Multi-model compatibilityIt supports mainstream large models such as OpenAI, Gemini, DeepSeek, and Doubao Seed, and allows for flexible switching of different underlying capabilities through a unified configuration interface.

Key information and usage requirements for DeerFlow 2.0

• Project Background: ByteDance's open-source super intelligent agent framework, licensed under the MIT license.
• Release timeVersion 2.0 was officially released on February 28, 2026, and immediately topped the GitHub Trending charts on the same day.
• Core positioningThe deep research tool has been fully upgraded from version 1.0 to the general-purpose intelligent agent runtime (Super Agent Harness).
• Task capabilityIt supports handling complex, multi-step tasks ranging from minutes to hours, covering various scenarios such as research, coding, and creation.
• Python versionRequires Python 3.12 or later to be installed on the system.
• Node.js versionRequires Node.js version 22 or higher to be installed on the system.
• Package ManagerYou need to install pnpm and uv as front-end and back-end package management tools.

How to use DeerFlow 2.0

• Cloning repository:implement git clone https://github.com/bytedance/deer-flow.git Clone the project to your local machine and navigate to the project directory. cd deer-flow.
• Generate configuration:run make config Automatically generated config.yaml and .env Configuration file.
• Configuration Model:edit config.yaml In the file, add the required models (such as GPT-4, Gemini, DeepSeek, Doubao, etc.), and set parameters such as model name, API type, and key variables.
• Set key:exist .env Enter the API keys of each service provider in the file, such as... OPENAI_API_KEY=your-key.
• Select deployment method
  • Docker method:implement make docker-init Pull the sandbox image and then execute. make docker-start Start the service.
  • Local method:implement make check Check dependencies before execution. make install Install dependencies, and finally execute. make dev Start the service.
• Access usageOpen your browser and visit http://localhost:2026 to enter the web interface. Enter the task command to start using it.
• IM channel access:exist config.yaml Enable Lark, Telegram, or Slack configurations, set the App ID, key, and other parameters for the corresponding platform, and you can interact with DeerFlow via commands in the chat software.

DeerFlow 2.0 project address

• Project official websitehttps://deerflow.tech
• GitHub repositoryhttps://github.com/bytedance/deer-flow

Comparison of DeerFlow 2.0 with similar competing products

Application Scenarios of DeerFlow 2.0

• In-depth researchIt automatically collects information from multiple sources and generates research reports, making it suitable for complex information processing tasks such as academic research, competitor analysis, and industry surveys.
• Web developmentIt can generate complete and deliverable website pages with one click, automating the entire process from UI design to front-end code, such as football league official websites and 3D interactive weather interfaces.
• Content creationTransform complex concepts into child-friendly educational comics, or automatically generate multimedia materials such as PPTs, podcast scripts, and video content.
• Data AnalysisIt automatically performs data-driven tasks such as data cleaning, visualization chart generation, and business intelligence report writing.