Deep Research
with DeerFlow
An open-source SuperAgent harness that researches, codes, and creates. With
the help of sandboxes, memories, tools, skills and subagents, it handles
different levels of tasks that could take minutes to hours.
Case Studies
See how DeerFlow is used in the wild
Agent Skills
Agent Skills are loaded progressively — only what's needed, when it's needed.
Extend DeerFlow with your own skill files, or use our built-in library.
Agent Runtime Environment
We give DeerFlow a "computer", which can execute commands, manage files, and run long tasks — all in a secure Docker-based sandbox
pygame==2.5.0
✔ Installed pygame
✔ Written 156 lines
✔ All sprites loaded
✔ Physics engine OK
✔ 60 FPS stable
Downloaded 12.4 MB
We recommend using All-in-One Sandbox that combines Browser, Shell, File, MCP and VSCode Server in a single Docker container.
IsolatedSafePersistentMountable FSLong-running
Whats New in DeerFlow 2.0
DeerFlow is now evolving from a Deep Research agent into a full-stack Super Agent
Long/Short-term Memory
Now the agent can better understand you
Planning and Sub-tasking
Plans ahead, reasons through complexity, then executes sequentially or in parallel
Skills and Tools
Plug, play, or even swap built-in tools. Build the agent you want.
Sandbox with File System
Read, write, run — like a real computer
Multi-Model Support
Doubao, DeepSeek, OpenAI, Gemini, etc.
Open Source
MIT License, self-hosted, full control
Join the Community
Contribute brilliant ideas to shape the future of DeerFlow. Collaborate, innovate, and make impacts.