ScriptsApr 6, 2026·2 min read

DeerFlow — SuperAgent for Research, Code & Creative

ByteDance's open-source long-horizon SuperAgent with sub-agents, sandboxes, memory, and skills. Handles complex tasks spanning minutes to hours. 58,300+ stars, MIT license.

SC
Script Depot · Community
Quick Use

Use it first, then decide how deep to go

This block should tell both the user and the agent what to copy, install, and apply first.

git clone https://github.com/bytedance/deer-flow.git && cd deer-flow
make config       # Generate config from templates
make docker-start # Launch via Docker (recommended)

Set API keys in .env: OPENAI_API_KEY, TAVILY_API_KEY, etc. Access the UI at http://localhost:2026.

Intro

DeerFlow (Deep Exploration and Efficient Research Flow) is ByteDance's open-source SuperAgent framework with 58,300+ GitHub stars and MIT license. It orchestrates sub-agents, sandboxes, memory systems, and extensible skills to handle complex, long-horizon tasks — from deep research and code generation to creative production. Hit #1 on GitHub Trending when v2.0 launched in February 2026. Supports Docker/Kubernetes deployment, multiple LLM providers (OpenAI, Anthropic, DeepSeek, Doubao), and integrations with Telegram, Slack, Feishu/Lark.

Best for: developers and researchers who need autonomous agents for complex, multi-step tasks lasting minutes to hours. Works with: OpenAI, Anthropic Claude, DeepSeek, OpenRouter, Doubao, Kimi. Setup time: under 5 minutes with Docker.

DeerFlow — Architecture & Features

Core Components

Component Purpose
Sub-Agents Hierarchical agent orchestration — delegate and specialize
Sandboxes Docker/local/K8s isolation for safe code execution
Memory System Persistent context across sessions
Skills Composable tools: web search, code execution, file ops
Context Engineering Sophisticated prompt management and state handling
Multi-Channel Telegram, Slack, Feishu/Lark, WeCom

Deployment Modes

Mode Description
Standard Separate LangGraph server + frontend
Gateway Embedded agent runtime — fewer processes, lower overhead
Docker make docker-start — recommended for production
Kubernetes K8s-ready with sandbox isolation

Tech Stack

  • Backend: Python 3.12+, LangChain, LangGraph, FastAPI
  • Frontend: Node.js 22+, React/TypeScript
  • Observability: LangSmith/Langfuse tracing
  • Package Management: uv (Python), pnpm (Node.js)

Supported LLM Providers

OpenAI, Anthropic Claude, OpenRouter, DeepSeek, Doubao, Kimi — configure in config.yaml.

FAQ

Q: What is DeerFlow? A: ByteDance's open-source SuperAgent framework that orchestrates sub-agents, sandboxes, and memory for complex tasks spanning minutes to hours. 58,300+ stars, MIT license.

Q: Is it free? A: Yes, MIT license. You provide your own LLM API keys.

Q: How does it compare to CrewAI or AutoGen? A: DeerFlow focuses on long-horizon tasks with sandbox execution, persistent memory, and multi-channel integration. CrewAI/AutoGen focus more on multi-agent conversation patterns.

🙏

Source & Thanks

Created by ByteDance. Licensed under MIT.

deer-flow — ⭐ 58,300+

Discussion

Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.

Related Assets

Pydantic — Data Validation for AI Agent Pipelines

Python's most popular data validation library, essential for AI agent tool definitions. Pydantic enforces type safety in LLM structured outputs, API schemas, and config files.

Script Depot

Open WebUI — Self-Hosted ChatGPT Alternative

Feature-rich open-source web UI for running local and remote LLMs. Open WebUI supports Ollama, OpenAI, Claude API with RAG, tools, multi-user, and mobile-friendly interface.

Script Depot

Docusaurus — Build AI Tool Documentation Sites

Meta's documentation framework used by major AI projects. Docusaurus creates fast, SEO-friendly docs sites with MDX, versioning, i18n, and Algolia search built in.

Script Depot
◈Home🔍Search👤Me