AutoGPT vs deer-flow
Two of the top projects in Agent platforms, side by side: who leads on each score part and who each one suits, then setup, license, activity and what each review found.
AutoGPT
Role-based AI agents coordinated by one lead agent, plus a visual workflow builder
deer-flow
Agent harness running sub-agents, memory and sandboxes with pluggable skills
Verdict
AutoGPT ranks higher, 95 to 93. Of the 4 score parts we measure for both, AutoGPT leads on 2, deer-flow on 1 and 1 is even.
- AdoptionAutoGPT leads by 6 points, 100 to 94.
- FreshnessEven at 100.
- Maintenancedeer-flow leads by 5 points, 93 to 88.
- Agent-readyAutoGPT leads by 15 points, 85 to 70: a wide gap.
Pick AutoGPT if you are
Teams delegating content, sales and recurring work to AI agents
- Single-container Docker image for self-hosting, plus an installer for Linux, macOS and Windows
- Visual builder with branching logic and a marketplace of reusable workflows
Pick deer-flow if you are
Developers self-hosting a research and task agent with sub-agents and sandboxing
- Interactive `make setup` wizard and `make doctor` health check for configuration
- Many search and fetch backends: Tavily, Brave, DuckDuckGo, SearXNG, Jina, InfoQuest
| What we compare | AutoGPT | deer-flow |
|---|---|---|
| Score parts, out of 100 | ||
| Adoption | 100, widely used | 94, widely used |
| Momentum | Not scored for this kind | Not scored for this kind |
| Freshness | 100, active | 100, active |
| Maintenance | 88, healthy | 93, healthy |
| Agent-ready | 85, ready | 70, partly |
| Facts from GitHub and the README | ||
| Stars | 188k | 85.2k |
| License | custom license (read the license) | MIT (permissive) |
| Last commit | Nov 2026 | Nov 2026 |
| Last release | Nov 2026 | Nov 2026 |
| Language | Python | Python |
| Docker | Yes | Yes |
| GPU | Not stated | Not stated |
| arm64 or Apple Silicon | Mentioned | Not stated |
AutoGPT
AutoGPT is a platform where a coordinator agent (Otto) hands work to configurable specialist agents called Experts, each with its own instructions, skills and connected apps. It also has a block-based visual workflow builder and a workflow marketplace. It runs as a paid hosted service or self-hosted from a single Docker image with your own model credentials.
Who it is for: Teams delegating content, sales and recurring work to AI agents
Strengths
- Single-container Docker image for self-hosting, plus an installer for Linux, macOS and Windows
- Visual builder with branching logic and a marketplace of reusable workflows
- Per-Expert account access and spending controls
- Self-hosting accepts your own model credentials or a local model
Weaknesses
- Platform code is Polyform Shield, not OSI open source; only classic/ is MIT
- Hosted Platform is paid; built-in model access is not included when self-hosting
- README gives no RAM or GPU requirements; they are in the linked guide
- Original standalone agent moved to classic/; main README centers on the platform
- Needs Docker, model provider credentials or a local model
- Models: bring-your-own model credentials, local models
- port 3000
deer-flow
DeerFlow 2.0 is a ground-up rewrite of ByteDance's deep research framework. It is a Python backend with a web UI that orchestrates a lead agent, sub-agents, long-term memory and sandboxed execution, extended through skills and MCP. Setup runs through a `make setup` wizard that writes config.yaml and .env.
Who it is for: Developers self-hosting a research and task agent with sub-agents and sandboxing
Strengths
- Interactive `make setup` wizard and `make doctor` health check for configuration
- Many search and fetch backends: Tavily, Brave, DuckDuckGo, SearXNG, Jina, InfoQuest
- Optional tracing through LangSmith, Langfuse and Monocle
- Includes sandbox mode, MCP, IM channels, scheduled tasks and an embedded Python client
Weaknesses
- 2.0 shares no code with v1; the original research framework lives on the 1.x branch
- README is dense with option-level detail, which buries the core setup steps
- Gateway admin is equivalent to code execution, so deployment needs care
- README recommends specific hosted models; local model guidance is not in the visible text
- Docker + Compose
- Needs LLM provider API key, Python 3.12+, Node.js 22+, Docker (recommended)
- Models: OpenAI-compatible endpoints, Gemini, Claude, Codex, Doubao-Seed-2.0-Code
More in Agent platforms
Building with these?
We help teams pick an open-source AI stack and run it in production. Client work never changes a rank.