$ show OpenBMB/ChatDev
ChatDev 2.0 (DevAll) is a zero-code multi-agent orchestration platform for building and executing graph-based AI workflows. Users define directed acyclic or cyclic graphs in YAML, composing agent nodes (with LLM providers, memory, tools, thinking modes), human-in-the-loop nodes, Python executors, subgraphs, and control-flow edges with conditions and data transformations. The system supports dynamic parallel execution patterns (Map/Tree), multiple memory stores (simple, file-based FAISS, cloud Mem0), and extensible tooling via local Python functions or MCP servers. The architecture is a full-stack application: a FastAPI backend with REST and WebSocket endpoints for workflow execution, batch runs, artifact streaming, and session management; a Vue 3 frontend with a visual workflow editor (VueFlow), dynamic schema-driven forms, and real-time execution monitoring; a CLI entry point (`run.py`); and a Python SDK (`runtime/sdk.py`) for programmatic use. The codebase is well-structured with registry patterns for extensibility, declarative configuration dataclasses with auto-generated UI forms, and a modular runtime with pluggable node executors, edge conditions, and memory stores. While primarily an application, the project has strong research roots (OpenBMB lab, Tsinghua NLP) and includes academic references (ChatDev, Co-Learning, MacNet). It serves as both a production-grade tool for multi-agent automation and a research platform for experimenting with agent architectures, memory systems, and collaborative workflows. The repository includes 20+ example workflows demonstrating diverse use cases from software development and data visualization to 3D scene generation and deep research.
The LLM identified 5 files worth highlighting for reuse:
Generic lazy-registration pattern used across the codebase for node types, edge conditions, memory stores, thinking modes, and tooling types. Provides a clean, thread-safe way to register and retrieve implementations by name, enabling plugin-like extensibility without modifying core code. Could be extracted as a standalone utility for any project needing a registry.
BaseConfig class with FIELD_SPECS metadata that drives automatic UI form generation and schema export. Each config subclass declares field types, defaults, validation rules, and child routes. This pattern decouples configuration from rendering and enables the frontend to dynamically build forms for any node or edge type. Reusable for any project needing schema-driven configuration.
File-based memory store using FAISS for vector indexing and semantic search. Supports incremental updates, chunking with overlap and sentence boundary detection, metadata tracking, and embedding configuration. Demonstrates a clean abstraction for persistent agent memory that could be reused in other agent frameworks or RAG systems.
Implements Map (fan-out) and Tree (fan-out with reduce) dynamic execution patterns on workflow edges. Uses thread pools for parallel node execution and supports split strategies (message, regex, JSON path). This pattern is reusable for any dataflow pipeline requiring parallel processing and aggregation, independent of the agent workflow context.
A Vue 3 component that recursively renders configuration forms based on a backend-provided schema. Supports nested objects, arrays, enums, booleans, strings, numbers, and conditional child routes. Demonstrates a robust pattern for building dynamic UIs that adapt to evolving configuration schemas without hardcoding form layouts. Could be extracted as a reusable library for any schema-driven application.