DeerFlow: Open-Source Long-Horizon SuperAgent
Overview
DeerFlow (formerly deer-flow) has emerged as one of the fastest-growing open-source AI agent projects, reaching 75,560 GitHub stars. The project describes itself as an "open-source long-horizon SuperAgent harness that researches, codes, and creates" — capable of handling complex, multi-step tasks that require sustained reasoning and execution over extended periods.
Unlike traditional AI agents that operate in short, stateless interactions, DeerFlow is designed for long-horizon tasks — projects that may require hours or days of autonomous work, with persistent memory, sandboxed execution environments, and the ability to spawn sub-agents for parallel work.
Key Features
- Long-Horizon Execution: Maintains context and progress across extended task sessions
- Sandboxed Environments: Runs code and experiments in isolated containers for safety
- Persistent Memory: Remembers past work and builds on previous results
- Sub-Agent Orchestration: Spawns and manages specialized sub-agents for parallel tasks
- Tool Ecosystem: Built-in tools for research, coding, testing, and content creation
Why It Matters
DeerFlow represents a new generation of AI agent frameworks that are designed for autonomous project execution rather than simple task completion. This shift from "task agents" to "project agents" has significant implications for:
- Software Development: Agents that can plan, code, test, and iterate on entire features
- Research: Agents that can conduct literature reviews, design experiments, and analyze results
- Content Production: Agents that can research, write, edit, and publish long-form content
Getting Started
# Clone the repository
git clone https://github.com/bytedance/deer-flow.git
cd deer-flow
# Install dependencies
pip install -r requirements.txt
# Configure your API keys
cp .env.example .env
# Edit .env with your API keys
# Run DeerFlow
python deer_flow.py --task "Your long-horizon task description"