An open-source agent harness that orchestrates sub-agents, long-term memory, and sandboxed execution to perform complex, long-horizon tasks such as research, coding, and content creation.
Project overview
An open-source agent harness that orchestrates sub-agents, long-term memory, and sandboxed execution to perform complex, long-horizon tasks such as research, coding, and content creation.
Project type
AI Agent · Workflow · Infrastructure
Use cases
Search & Research · Coding & Development · Automation
Deployment
Refer to project documentation
License
MIT
Best for
Developers, AI engineers, and enterprise teams building automated workflows for deep research or coding tasks.
Key capabilities
Orchestrates sub-agents, memory, and sandboxes to perform complex long-horizon tasks.
Supports Local Execution, Docker Execution, and Docker Execution with Kubernetes for sandbox code.
Supports configurable MCP servers and skills to extend its capabilities, including HTTP/SSE and stdio MCP servers.
Supports receiving tasks from messaging apps like Telegram, Slack, Feishu, WeChat, WeCom, and DingTalk.
Ships a skill that lets users interact with a running instance directly from Claude Code.
Maintains long-term memory for the agent.
Built-in integrations for LangSmith, Langfuse, and Monocle tracing.
Limitations and risks
Deploying in untrusted environments may introduce security risks such as unauthorized illegal invocation.
Compliance and legal risks exist if illegally invoked for cyberattacks or data theft.
Getting started
Requires Python 3.12+ and Node.js 22+. Users clone the repository and run make setup, make install, and make dev to configure LLM providers and start the application.
Evidence and sources
GitHub project description: An open-source long-horizon SuperAgent harness that researches, codes, and creates.
README: DeerFlow (**D**eep **E**xploration and **E**fficient **R**esearch **Flow**) is an open-source **super agent harness** that orchestrates **sub-agents**, **memory**, and **sandboxes…
README: DeerFlow 2.0 is a ground-up rewrite.
README: DeerFlow supports multiple sandbox execution modes: - **Local Execution** (runs sandbox code directly on the host machine) - **Docker Execution** (runs sandbox code in isolated Do…
README: DeerFlow supports configurable MCP servers and skills to extend its capabilities.