A multi-agent orchestration platform focused on dynamic, autonomous organization of agent fleets.
Flock Ops is a fork of ChatDev 2.0 (DevAll) by OpenBMB. Where ChatDev models agent collaboration around human-driven organizational patterns -- roles like CEO, CTO, and Programmer participating in structured seminars -- Flock Ops takes a different approach: agents are dynamically composed, sequenced, and coordinated as an autonomous fleet, without relying on fixed human hierarchies.
- OS: macOS / Linux / WSL / Windows
- Python: 3.12+
- Node.js: 18+
-
Backend Dependencies:
pip install -r requirements.txt
-
Frontend Dependencies (Vite + Vue 3):
cd frontend && npm install
- Config file:
cp config.example.yaml config.yaml
- Provider credentials: Set
provider.api_keyandprovider.base_urlinconfig.yamlfor your LLM provider. - YAML placeholders: Use
${VAR}(e.g.,${API_KEY}) in workflow files to reference provider credentials.
Start both backend and frontend:
python dev.py devThen access the Web Console at http://localhost:5173.
-
Start Backend:
python server_main.py --reload
Remove
--reloadif output files (e.g., GameDev) trigger restarts, which interrupts tasks and loses progress. -
Start Frontend:
cd frontend && npm run dev
Then access the Web Console at http://localhost:5173.
Tip: If the frontend fails to connect to the backend, the default port
6400may already be occupied. Switch both services to an available port:- Backend: start with
--port 6401 - Frontend: set
VITE_API_BASE_URL=http://localhost:6401
- Backend: start with
-
Help command:
make help -
Sync YAML workflows to frontend:
make sync
Uploads all workflow files from
yaml_instance/to the database. -
Validate all YAML workflows:
make validate-yamls
Checks all YAML files for syntax and schema errors.
Alternatively, run the entire application using Docker Compose.
-
Prerequisites:
- Docker and Docker Compose installed.
- Ensure you have a
.envfile in the project root for your API keys.
-
Build and Run:
docker compose up --build
-
Access:
- Backend:
http://localhost:6400 - Frontend:
http://localhost:5173
- Backend:
The services will automatically restart if they crash, and local file changes will be reflected inside the containers for live development.
The Flock Ops interface provides a seamless experience for both construction and execution.
-
Workflow: A visual canvas to design your multi-agent systems. Configure node parameters, define context flows, and orchestrate agent interactions with drag-and-drop ease.
-
Launch: Initiate workflows, monitor real-time logs, inspect intermediate artifacts, and provide human-in-the-loop feedback.
For automation and batch processing, use the Python SDK to execute workflows programmatically and retrieve results directly.
from runtime.sdk import run_workflow
result = run_workflow(
yaml_file="yaml_instance/demo.yaml",
task_prompt="Summarize the attached document in one sentence.",
attachments=["/path/to/document.pdf"],
variables={"API_KEY": "sk-xxxx"}
)
if result.final_message:
print(f"Output: {result.final_message.text_content()}")The project is organized into a modular structure:
- Core Systems:
server/hosts the FastAPI backend, whileruntime/manages agent abstraction and tool execution. - Orchestration:
workflow/handles the multi-agent logic, driven by configurations inentity/. - Frontend:
frontend/contains the Vue 3 Web Console. - Extensibility:
functions/is the place for custom Python tools.
Relevant reference documentation:
- Getting Started: Start Guide
- Core Modules: Workflow Authoring, Memory, and Tooling
Out-of-the-box templates for common scenarios are located in yaml_instance/.
- Demos: Files named
demo_*.yamlshowcase specific features or modules. - Implementations: Files named directly (e.g.,
ChatDev_v1.yaml) are full workflow implementations.
| Category | Workflow | Description |
|---|---|---|
| Data Visualization | data_visualization_basic.yaml, data_visualization_enhanced.yaml |
Generate charts from datasets |
| 3D Generation | blender_3d_builder_simple.yaml, blender_3d_builder_hub.yaml, blender_scientific_illustration.yaml |
3D model generation (requires Blender and blender-mcp) |
| Game Dev | GameDev_v1.yaml, ChatDev_v1.yaml |
Game design and development |
| Deep Research | deep_research_v1.yaml |
Automated research workflows |
Flock Ops is forked from ChatDev by OpenBMB / THUNLP. The original research papers:
@article{chatdev,
title = {ChatDev: Communicative Agents for Software Development},
author = {Chen Qian and Wei Liu and Hongzhang Liu and Nuo Chen and Yufan Dang and Jiahao Li and Cheng Yang and Weize Chen and Yusheng Su and Xin Cong and Juyuan Xu and Dahai Li and Zhiyuan Liu and Maosong Sun},
journal = {arXiv preprint arXiv:2307.07924},
url = {https://arxiv.org/abs/2307.07924},
year = {2023}
}
@article{puppeteer,
title={Multi-Agent Collaboration via Evolving Orchestration},
author={Yufan Dang and Chen Qian and Xueheng Luo and Jingru Fan and Zihao Xie and Ruijie Shi and Weize Chen and Cheng Yang and Xiaoyin Che and Ye Tian and Xuantang Xiong and Lei Han and Zhiyuan Liu and Maosong Sun},
journal={arXiv preprint arXiv:2505.19591},
url={https://arxiv.org/abs/2505.19591},
year={2025}
}