
DeerFlow
β Text source-checked Nov 7, 2026An open-source SuperAgent harness that researches, codes, and creates.
Source-checked against the vendor's site Β· rebuilt from the vendor's own pagesProfile updated Link checked How we rate
In short
DeerFlow is a free, open-source SuperAgent harness for deep research and coding. It uses multi-model support and secure sandboxes to handle complex, long-running tasks.
DeerFlow is a framework for building and operating agent systems. The system provides a runtime harness for composing agents with memory, tools, skills, sandboxes, and subagents, along with an application layer that turns those capabilities into a usable product experience.
DeerFlow handles different levels of tasks that could take minutes to hours using sandboxes, memories, tools, skills, and subagents. The product provides a secure Docker-based sandbox environment that can execute commands, manage files, and run long tasks.
Agent skills are loaded progressively, ensuring only necessary components are active when needed. The framework is evolving from a deep research agent into a full-stack Super Agent.
According to DeerFlow: DeerFlow is a framework for building and operating agent systems. Its pages list this strength: It includes a secure Docker-based sandbox for command execution and file management.
β The AI Tool Bible editorial team
Pros
- β It includes a secure Docker-based sandbox for command execution and file management.
- β It supports long-running tasks that can take minutes to hours.
- β It features progressive loading of agent skills.
Facts
- Platforms
- web, api, vscode-extension (not source-checked)
- API available
- Yes (not source-checked)
- Open source
- Yes (license: Apache-2.0)
- GitHub stars
- 6,333 (checked Nov 13, 2026) (not source-checked)
- Last GitHub push
- Sep 14, 2026 (not source-checked)
- First commit
- Aug 2025 (not source-checked)
- Model used
- Multi-model (Doubao, DeepSeek, OpenAI, Gemini) (not source-checked)
- Package downloads (30 d)
- 5,680 Β· npm @agent-infra/sandbox Β· checked Nov 10, 2026 (not source-checked)
- Hacker News
- 2 stories Β· top 1 points Β· last Mar 25, 2026 Β· checked Nov 10, 2026
- Earliest Wayback capture
- 2025-05-11 (earliest archived capture of the site, not a launch date)
Where this tool sits
Use cases
Frequently asked
- How much does DeerFlow cost?
- DeerFlow is free. The provided documentation does not list specific pricing tiers or additional costs, describing it simply as a free, open-source framework.
- Which AI models does DeerFlow support?
- DeerFlow supports a multi-model approach, including Doubao, DeepSeek, OpenAI, and Gemini. This allows the agent system to leverage different underlying models for various tasks.
- Is DeerFlow safe for executing code?
- Yes, the product provides a secure Docker-based sandbox environment. This sandbox can execute commands, manage files, and run long tasks safely, isolating operations from the main system.
- What tasks is DeerFlow best suited for?
- DeerFlow is designed for deep research and data analysis. It handles tasks ranging from minutes to hours in duration by utilizing sandboxes, memories, tools, skills, and subagents.
- How does DeerFlow manage agent skills?
- Agent skills are loaded progressively. This ensures that only the necessary components are active when needed, optimizing performance and resource usage during task execution.