### Install Backend Dependencies

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Installs Python backend dependencies using uv. Ensure Python 3.12+ is installed.

```bash
uv sync
```

--------------------------------

### Build and Run Docker Compose

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Builds and starts the ChatDev application using Docker Compose. Ensure Docker and Docker Compose are installed and a .env file is present.

```bash
# From the project root
docker compose up --build
```

--------------------------------

### Start Backend Manually

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Starts the ChatDev backend server. Use --reload to watch for source file changes. Customize reload directories with --reload-dir or --reload-exclude.

```bash
# Run from the project root
uv run python server_main.py --port 6400 --reload
```

--------------------------------

### Install Frontend Dependencies

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Installs Node.js frontend dependencies using npm. Navigate to the frontend directory first.

```bash
cd frontend && npm install
```

--------------------------------

### Run Application with Makefile

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Starts both the backend and frontend services simultaneously using the Makefile. Access the web console at http://localhost:5173.

```bash
make dev
```

--------------------------------

### Start Frontend Manually

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Starts the ChatDev frontend development server. Ensure VITE_API_BASE_URL points to the backend. Access the web console at http://localhost:5173.

```bash
cd frontend
VITE_API_BASE_URL=http://localhost:6400 npm run dev
```

--------------------------------

### Configure Environment Variables

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Copies the example environment file to be used for configuration. Set API keys and base URLs in the .env file.

```bash
cp .env.example .env
```

--------------------------------

### Subgraph Node Examples

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/subgraph.md

Illustrative examples of how to use the Subgraph node, including referencing external YAML files and defining subgraphs inline.

```APIDOC
## Subgraph Node Examples

### Referencing External File

```yaml
nodes:
  - id: Review Process
    type: subgraph
    config:
      type: file
      config:
        path: common/review_flow.yaml
```

### Inline Subgraph Definition

```yaml
nodes:
  - id: Translation Unit
    type: subgraph
    config:
      type: config
      config:
        id: translation_subgraph
        description: Multi-language translation subprocess
        nodes:
          - id: Translator
            type: agent
            config:
              provider: openai
              name: gpt-4o
              role: You are a professional translator who translates content to the target language.
          - id: Proofreader
            type: agent
            config:
              provider: openai
              name: gpt-4o
              role: You are a proofreading expert who checks and polishes translated content.
        edges:
          - from: Translator
            to: Proofreader
        start: [Translator]
        end: [Proofreader]
```

### Combining Multiple Subgraphs

```yaml
nodes:
  - id: Input Handler
    type: agent
    config:
      provider: openai
      name: gpt-4o

  - id: Analysis Module
    type: subgraph
    config:
      type: file
      config:
        path: modules/analysis.yaml

  - id: Report Module
    type: subgraph
    config:
      type: file
      config:
        path: modules/report_gen.yaml

edges:
  - from: Input Handler
    to: Analysis Module
  - from: Analysis Module
    to: Report Module
```
```

--------------------------------

### Passthrough Node Basic Usage Example

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/passthrough.md

A minimal example demonstrating the basic usage of a Passthrough node within a workflow.

```yaml
nodes:
  - id: Router
    type: passthrough
    config: {}
```

--------------------------------

### Literal Node Examples

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/literal.md

Illustrative examples of how to use the Literal node in various scenarios.

```APIDOC
## Literal Node Examples

### Basic Usage
Outputs a welcome message with the assistant role.
```yaml
nodes:
  - id: Welcome Message
    type: literal
    config:
      content: |
        Welcome to the intelligent assistant! Please describe your needs.
      role: assistant
```

### Injecting Fixed Context
Injects predefined rules or context as a user message.
```yaml
nodes:
  - id: Context Injector
    type: literal
    config:
      content: |
        Please note the following rules:
        1. Answers must be concise and clear
        2. Reply in English
        3. If uncertain, please state so
      role: user

  - id: Assistant
    type: agent
    config:
      provider: openai
      name: gpt-4o

edges:
  - from: Context Injector
    to: Assistant
```

### Fixed Responses in Conditional Branches
Provides fixed responses for different conditions determined by a classifier.
```yaml
nodes:
  - id: Classifier
    type: agent
    config:
      provider: openai
      name: gpt-4o
      role: Determine user intent, reply with KNOWN or UNKNOWN

  - id: Known Response
    type: literal
    config:
      content: I can help you complete this task.
      role: assistant

  - id: Unknown Response
    type: literal
    config:
      content: Sorry, I cannot understand your request. Please describe it in a different way.
      role: assistant

edges:
  - from: Classifier
    to: Known Response
    condition:
      type: keyword
      config:
        any: [KNOWN]
  - from: Classifier
    to: Unknown Response
    condition:
      type: keyword
      config:
        any: [UNKNOWN]
```

### Testing Purposes
Provides a multi-line test input for downstream processing nodes.
```yaml
nodes:
  - id: Test Input
    type: literal
    config:
      content: |
        This is a test text for verifying downstream processing logic.
        Contains multiple lines.
      role: user

  - id: Processor
    type: python
    config:
      timeout_seconds: 30

edges:
  - from: Test Input
    to: Processor

start: [Test Input]
```
```

--------------------------------

### Define a Read Text File Tool

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/function.md

Example implementation of a function tool using type hints and ParamMeta for documentation.

```python
from typing import Annotated
from utils.function_catalog import ParamMeta


def read_text_file(
    path: Annotated[str, ParamMeta(description="workspace-relative path")],
    *,
    encoding: str = "utf-8",
    _context: dict | None = None,
) -> str:
    ctx = FileToolContext(_context)
    target = ctx.resolve_under_workspace(path)
    return target.read_text(encoding=encoding)
```

--------------------------------

### FastMCP Sample Server Implementation

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/mcp.md

A Python example using the FastMCP library to create a simple MCP server with a `rand_num` tool. This script can be run locally or remotely.

```python
from fastmcp import FastMCP
import random

mcp = FastMCP("Company Simple MCP Server", debug=True)

@mcp.tool
def rand_num(a: int, b: int) -> int:
    return random.randint(a, b)

if __name__ == "__main__":
    mcp.run()
```

--------------------------------

### Loop Counter Node Usage Example

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/loop_counter.md

An example demonstrating how to configure and use the Loop Counter node in a workflow.

```APIDOC
## Loop Counter Node Example

### Description
This example shows a basic implementation of the Loop Counter node to limit a workflow to 5 iterations.

### Request Example
```yaml
nodes:
  - id: Iteration Guard
    type: loop_counter
    config:
      max_iterations: 5
      reset_on_emit: true
      message: Maximum iteration count reached, process terminated.
```
```

--------------------------------

### Configure a Loop Timer Node

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/loop_timer.md

Example configuration for a Loop Timer node set to a 5-minute duration.

```yaml
nodes:
  - id: Time Guard
    type: loop_timer
    config:
      max_duration: 5
      duration_unit: minutes
      reset_on_emit: true
      message: Time limit reached (5 minutes), process terminated.
```

--------------------------------

### Configure Agent Tooling

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/README.md

Example YAML configuration for an agent node using function-based tooling with auto-loading enabled.

```yaml
nodes:
  - id: solve
    type: agent
    config:
      provider: openai
      model: gpt-4o-mini
      prompt_template: solver
      tooling:
        type: function
        config:
          tools:
            - name: describe_available_files
            - name: load_file
          auto_load: true
          timeout: 20
```

--------------------------------

### Configure Agent Tooling in YAML

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/function.md

Example YAML configuration for enabling specific function tools within an agent node.

```yaml
nodes:
  - id: summarize
    type: agent
    config:
      tooling:
        type: function
        config:
          tools:
            - name: describe_available_files
            - name: read_text_file
```

--------------------------------

### Travel Planning Workflow with Map and Tree Combination

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/dynamic_execution.md

An example of a complex workflow combining 'map' for parallel planning requests and 'tree' for aggregating results. This setup is suitable for scenarios where multiple independent planning tasks need to be executed and their outcomes consolidated into a single coherent plan.

```yaml
graph:
  nodes:
    - id: Eat Planner
      type: literal
      config:
        content: Plan what to eat in Shanghai
        role: user
    - id: Play Planner
      type: literal
      config:
        content: Plan what to do in Shanghai
        role: user
    - id: Stay Planner
      type: literal
      config:
        content: Plan where to stay in Shanghai
        role: user
    - id: Collector
      type: passthrough
      config:
        only_last_message: false
    - id: Travel Executor
      type: agent
      config:
        name: gpt-4o
        role: You are a travel planner. Please plan according to user requests.
    - id: Final Aggregator
      type: agent
      config:
        name: gpt-4o
        role: Please integrate the inputs into a complete travel plan.

  edges:
    - from: Eat Planner
      to: Collector
    - from: Play Planner
      to: Collector
    - from: Stay Planner
      to: Collector
    - from: Collector
      to: Travel Executor
      dynamic:                  # Map fan-out: 3 planning requests → 3 parallel executions
        type: map
        split:
          type: message
        config:
          max_parallel: 10
    - from: Travel Executor
      to: Final Aggregator
      dynamic:                  # Tree reduce: 3 results → 1 final plan
        type: tree
        split:
          type: message
        config:
          group_size: 2
          max_parallel: 10
```

--------------------------------

### Local MCP Configuration Example

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/mcp.md

Configure local MCP agents using this YAML, specifying the command, arguments, and working directory. The `wait_for_log` field is crucial for signaling readiness.

```yaml
nodes:
  - id: local_mcp
    type: agent
    config:
      tooling:
        type: mcp_local
        config:
          command: uvx
          args:
            - blender-mcp
          cwd: ${REPO_ROOT}
          wait_for_log: "MCP ready"
          startup_timeout: 8
```

--------------------------------

### Install ChatDev Skills for OpenClaw

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Installs the necessary ChatDev skills for integration with OpenClaw. This command should be run within your OpenClaw instance.

```bash
clawdhub install chatdev
```

--------------------------------

### Gemini Provider Configuration Example

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/workflow_authoring.md

Configure the Gemini provider with model details, API key, and specific parameters like response modalities and safety settings. Environment variables can be used for sensitive information like API keys.

```yaml
model:
  provider: gemini
  base_url: https://generativelanguage.googleapis.com
  api_key: ${GEMINI_API_KEY}
  name: gemini-2.0-flash-001
  input_mode: messages
  params:
    response_modalities: ["text", "image"]
    safety_settings:
      - category: HARM_CATEGORY_SEXUAL
        threshold: BLOCK_LOWER
```

--------------------------------

### Configure Loop Counter Node

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/loop_counter.md

Example configuration for a loop_counter node with a limit of 5 iterations and automatic reset.

```yaml
nodes:
  - id: Iteration Guard
    type: loop_counter
    config:
      max_iterations: 5
      reset_on_emit: true
      message: Maximum iteration count reached, process terminated.
```

--------------------------------

### Configure Thinking Module in YAML

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/thinking.md

Example of how to configure the 'reflection' thinking mode within a node's configuration in YAML. Ensure the API key is securely provided.

```yaml
nodes:
  - id: Thoughtful Agent
    type: agent
    config:
      provider: openai
      name: gpt-4o
      api_key: ${API_KEY}
      thinking:
        type: reflection
        config:
          reflection_prompt: |
            Please carefully review your response, considering:
            1. Is the logic sound?
            2. Are there any factual errors?
            3. Is the expression clear?
            Then provide an improved response.
```

--------------------------------

### Python Environment Management with uv

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/function_catalog.md

Tools for managing Python environments and dependencies using `uv`. This includes installing packages, initializing environments, and running modules or scripts.

```python
install_python_packages
init_python_env
uv_run
```

--------------------------------

### Human Node Prompt

Source: https://github.com/openbmb/chatdev/blob/main/frontend/public/tutorial-zh.md

This prompt provides instructions for the 'Human Node', guiding the user on how to provide feedback or accept the current output.

```text
请对文章给出修改建议，或输入 ACCEPT 跳出循环。
```

--------------------------------

### Typical Workflow with Python Node for Result Parsing

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/python.md

Illustrates a workflow where an LLM generates Python code, which is then parsed by a Python node. This setup is common for code generation tasks.

```yaml
nodes:
  - id: LLM Generator
    type: agent
    config:
      provider: openai
      name: gpt-4o
      api_key: ${API_KEY}
      role: You need to generate executable Python code based on user input. The code should be wrapped in ```python ```.

  - id: Result Parser
    type: python
    config:
      timeout_seconds: 30

edges:
  - from: LLM Generator
    to: Result Parser
```

--------------------------------

### Remote MCP Configuration Example

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/mcp.md

Use this YAML configuration for agents requiring remote MCP access via HTTP(S). Ensure the server endpoint and any necessary headers like Authorization are correctly set.

```yaml
nodes:
  - id: remote_mcp
    type: agent
    config:
      tooling:
        type: mcp_remote
        config:
          server: https://mcp.mycompany.com/mcp
          headers:
            Authorization: Bearer ${MY_MCP_TOKEN}
          timeout: 15
```

--------------------------------

### Launch FastMCP Server

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/mcp.md

Command to launch the FastMCP sample server using `uv` with HTTP transport. Adjust the port as needed.

```bash
uv run fastmcp run mcp_example/mcp_server.py --transport streamable-http --port 8010
```

--------------------------------

### Basic Python Node Configuration

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/python.md

Demonstrates basic configuration for a Python node, setting a custom timeout and environment variables.

```yaml
nodes:
  - id: Data Processor
    type: python
    config:
      timeout_seconds: 120
      env:
        key: value
```

--------------------------------

### Memory Store Configuration Sample

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/memory.md

Configure memory stores like 'simple' or 'file' with paths and embedding settings. Use environment variables for sensitive information like API keys.

```yaml
memory:
  - name: convo_cache
    type: simple
    config:
      memory_path: WareHouse/shared/simple.json
      embedding:
        provider: openai
        model: text-embedding-3-small
        api_key: ${API_KEY}
  - name: project_docs
    type: file
    config:
      index_path: WareHouse/index/project_docs.json
      file_sources:
        - path: docs/
          file_types: [".md", ".mdx"]
          recursive: true
      embedding:
        provider: openai
        model: text-embedding-3-small
```

--------------------------------

### Agent Node Memory Attachment Example

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/memory.md

Define memory attachments for an agent node, specifying which memories to use, retrieval stages, and read/write permissions. This example attaches 'convo_cache' for generation stages and 'project_docs' for reading.

```yaml
nodes:
  - id: answer
    type: agent
    config:
      provider: openai
      model: gpt-4o-mini
      prompt_template: answer_user
      memories:
        - name: convo_cache
          retrieve_stage: ["gen"]
          top_k: 5
          read: true
          write: true
        - name: project_docs
          read: true
          write: false
```

--------------------------------

### Basic Literal Node Usage

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/literal.md

Demonstrates the basic configuration of a Literal node to output a welcome message with the assistant role.

```yaml
nodes:
  - id: Welcome Message
    type: literal
    config:
      content: |
        Welcome to the intelligent assistant! Please describe your needs.
      role: assistant
```

--------------------------------

### Python File Reading Tool Example

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/zh/modules/tooling/function.md

Example of a Python function `read_text_file` that reads text from a file. It uses `Annotated` for parameter descriptions and accepts an optional `_context` dictionary for tool execution context. The `FileToolContext` is used to resolve file paths relative to the workspace.

```python
from typing import Annotated
from utils.function_catalog import ParamMeta


def read_text_file(
    path: Annotated[str, ParamMeta(description="workspace 相对路径")],
    *, 
    encoding: str = "utf-8",
    _context: dict | None = None,
) -> str:
    ctx = FileToolContext(_context)
    target = ctx.resolve_under_workspace(path)
    return target.read_text(encoding=encoding)
```

--------------------------------

### Launch Workflow Execution

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/ws_frontend_logic.md

Initiates the workflow execution by sending a POST request to the backend. It requires prerequisites like a selected file, task prompt or attachments, and a ready WebSocket connection.

```javascript
shouldGlow.value = false
status.value = 'Launching...'
// POST request to /api/workflow/execute with payload
// ...
status.value = 'Running...'
isWorkflowRunning.value = true
```

--------------------------------

### Agent Node Configuration with Function Tool Calling

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/agent.md

Set up an agent node to use function calling tools. Specify the tool type as 'function' and list the available tools. An optional execution timeout can be configured.

```yaml
nodes:
  - id: Assistant
    type: agent
    config:
      provider: openai
      name: gpt-4o
      api_key: ${API_KEY}
      tooling:
        type: function  # Tool type: function, mcp_remote, mcp_local
        config:
          tools:  # List of function tools from functions/function_calling/ directory
            - name: describe_available_files
            - name: load_file
          timeout: 20  # Optional: execution timeout (seconds)
```

--------------------------------

### Passthrough Node for Conditional Routing

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/passthrough.md

This example shows how a Passthrough node can be used in conjunction with conditional edges to route messages based on keywords.

```yaml
nodes:
  - id: Classifier
    type: agent
    config:
      provider: openai
      name: gpt-4o
      role: | 
        Classify input content, reply TECHNICAL or BUSINESS

  - id: Router
    type: passthrough
    config: {}

  - id: Tech Handler
    type: agent
    config:
      provider: openai
      name: gpt-4o

  - id: Biz Handler
    type: agent
    config:
      provider: openai
      name: gpt-4o

edges:
  - from: Classifier
    to: Router
  - from: Router
    to: Tech Handler
    condition:
      type: keyword
      config:
        any: [TECHNICAL]
  - from: Router
    to: Biz Handler
    condition:
      type: keyword
      config:
        any: [BUSINESS]
```

--------------------------------

### User Interaction Tool

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/function_catalog.md

Sends instructions to the user and receives a response, suitable for scenarios requiring human input.

```python
call_user
```

--------------------------------

### Mem0 Memory Configuration Sample

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/memory.md

Configure the 'mem0' memory store, which is cloud-managed and requires an API key and agent ID. No local embeddings or persistence are needed.

```yaml
memory:
  - name: agent_memory
    type: mem0
    config:
      api_key: ${MEM0_API_KEY}
      agent_id: my-agent
```

--------------------------------

### Registering a Custom Memory Store

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/memory.md

Demonstrates how to register a new memory store by subclassing MemoryBase and using the register_memory_store function. Ensure to update FIELD_SPECS and export the design template.

```python
register_memory_store("my_store", config_cls=..., factory=..., summary="...")
```

--------------------------------

### Utility Commands with Makefile

Source: https://github.com/openbmb/chatdev/blob/main/README.md

Provides access to various utility commands for managing the ChatDev application.

```bash
make help
```

```bash
make sync
```

```bash
make validate-yamls
```

--------------------------------

### Execute Workflow via CLI

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/workflow_authoring.md

Run a workflow instance using the command line interface.

```bash
python run.py --path yaml_instance/demo.yaml --name test_run
```

--------------------------------

### Basic Agent Node Configuration

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/agent.md

Configure a basic agent node with OpenAI provider, API endpoint, and a system role prompt. Ensure environment variables for BASE_URL and API_KEY are set.

```yaml
nodes:
  - id: Writer
    type: agent
    config:
      provider: openai
      base_url: ${BASE_URL}
      api_key: ${API_KEY}
      name: gpt-4o
      role: |
        You are a professional technical documentation writer. Please answer questions in clear and concise language.
      params:
        temperature: 0.7
        max_tokens: 2000
```

--------------------------------

### Establish WebSocket Connection

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/ws_frontend_logic.md

Establishes a new WebSocket connection. It first resets the connection state, then creates a new WebSocket instance, and sets up event handlers for connection open, message, error, and close events.

```javascript
const socket = new WebSocket('ws://localhost:8000/ws')
ws = socket
socket.onopen = () => {
  console.log('WebSocket connected')
}
socket.onmessage = (event) => {
  const data = JSON.parse(event.data)
  if (data.type === 'connection') {
    sessionId = data.session_id
    isConnectionReady.value = true
    shouldGlow.value = true
    status.value = 'Waiting for launch...'
  }
}
socket.onerror = (error) => {
  console.error('WebSocket error:', error)
}
socket.onclose = () => {
  console.log('WebSocket disconnected')
}
```

--------------------------------

### 配置 Agent 节点的 Thinking 模块

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/zh/modules/thinking.md

在 YAML 配置文件中通过 nodes[].config.thinking 声明思考模式及其配置。

```yaml
nodes:
  - id: Thoughtful Agent
    type: agent
    config:
      provider: openai
      name: gpt-4o
      api_key: ${API_KEY}
      thinking:
        type: reflection
        config:
          reflection_prompt: |
            请仔细审视你的回答，考虑以下方面：
            1. 逻辑是否严密
            2. 有无事实错误
            3. 表达是否清晰
            然后给出改进后的回答。
```

--------------------------------

### Agent Node with MCP Local Tooling

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/zh/nodes/agent.md

Configure an agent node to use MCP for local tool execution via standard input/output. Specify the command to launch the MCP server, its arguments, working directory, environment variables, and startup timeout.

```yaml
nodes:
  - id: Local MCP Agent
    type: agent
    config:
      provider: openai
      name: gpt-4o
      api_key: ${API_KEY}
      tooling:
        type: mcp_local
        config:
          command: uvx  # 启动命令
          args: ["mcp-server-sqlite", "--db-path", "data.db"]
          cwd: ${WORKSPACE}  # 可选，一般不需要配置
          env:  # 可选，一般不需要配置
            DEBUG: "true"
          startup_timeout: 10  # 可选：启动超时（秒）
```

--------------------------------

### Run ChatDev with Human-Agent-Interaction Mode

Source: https://github.com/openbmb/chatdev/blob/main/README.md

To activate Human-Agent-Interaction mode, use this command. You can play the role of a reviewer and make suggestions to the programmer.

```bash
python3 run.py --task [description_of_your_idea] --config "Human"
```

--------------------------------

### Basic Agent Node Configuration

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/zh/nodes/agent.md

Configure a basic agent node with a specific LLM provider, API endpoint, and system prompt. Ensure API keys and base URLs are correctly set, potentially using environment variables.

```yaml
nodes:
  - id: Writer
    type: agent
    config:
      provider: openai
      base_url: ${BASE_URL}
      api_key: ${API_KEY}
      name: gpt-4o
      role: |
        你是一位专业的技术文档撰写者，请用清晰简洁的语言回答问题。
      params:
        temperature: 0.7
        max_tokens: 2000
```

--------------------------------

### Injecting Fixed Context with Literal Node

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/literal.md

Shows how to use a Literal node to inject fixed context or instructions into a workflow, followed by an agent node.

```yaml
nodes:
  - id: Context Injector
    type: literal
    config:
      content: |
        Please note the following rules:
        1. Answers must be concise and clear
        2. Reply in English
        3. If uncertain, please state so
      role: user

  - id: Assistant
    type: agent
    config:
      provider: openai
      name: gpt-4o

edges:
  - from: Context Injector
    to: Assistant
```

--------------------------------

### Configure Regex Split Strategy

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/dynamic_execution.md

Use regular expressions to partition text content into parallel execution units.

```yaml
split:
  type: regex
  pattern: "(?s).{1,2000}(?:\\s|$)"  # Split every ~2000 characters
```

--------------------------------

### Implement a Human-Machine Collaboration Loop

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/human.md

Sets up an iterative loop between an agent and a human reviewer using conditional edges.

```yaml
nodes:
  - id: Article Writer
    type: agent
    config:
      provider: openai
      name: gpt-4o
      api_key: ${API_KEY}
      role: You are a professional writer who writes articles based on user requirements.

  - id: Human Reviewer
    type: human
    config:
      description: |
        Please review the article:
        - If satisfied with the result, enter ACCEPT to end the process
        - Otherwise, enter modification suggestions to continue iterating

edges:
  - from: Article Writer
    to: Human Reviewer
  - from: Human Reviewer
    to: Article Writer
    condition:
      type: keyword
      config:
        none: [ACCEPT]
        case_sensitive: false
```

--------------------------------

### Run ChatDev with Art Mode

Source: https://github.com/openbmb/chatdev/blob/main/README.md

To activate Art mode, use this command. This enables the designer agent to generate images used in the software.

```bash
python3 run.py --task [description_of_your_idea] --config "Art"
```

--------------------------------

### Configure a Basic Human Node

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/human.md

Defines a simple human review step within a workflow configuration.

```yaml
nodes:
  - id: Human Reviewer
    type: human
    config:
      description: Please review the above content. If satisfied, enter ACCEPT; otherwise, enter your modification suggestions.
```

--------------------------------

### Export Design Templates

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/workflow_authoring.md

Regenerate YAML templates and frontend mirror files after modifying configurations.

```bash
python -m tools.export_design_template \
  --output yaml_template/design.yaml \
  --mirror frontend/public/design_0.4.0.yaml
```

--------------------------------

### Import Tools via YAML

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/function_catalog.md

Reference tools in YAML by specifying their type and configuration. You can import entire modules or individual functions.

```yaml
tooling:
  - type: function
    config:
      tools:
        - name: file:All           # Import entire module
        - name: save_file          # Import single function
        - name: deep_research:All
```

--------------------------------

### User Interaction API

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/modules/tooling/function_catalog.md

Tool for sending instructions to the user.

```APIDOC
## User Interaction

### Description
Send instructions to the user and get a response.

### Functions
- **call_user**: Send instructions to the user
```

--------------------------------

### Editor 2 Prompt Configuration

Source: https://github.com/openbmb/chatdev/blob/main/frontend/public/tutorial-en.md

Use this prompt to configure the Editor 2 agent for integrating and polishing articles based on revision suggestions. Ensure the output is only the revised article.

```text
You are an editor skilled at integrating and polishing an article.
Based on the article and the revision suggestions, refine and revise the article. Output only the revised article.
```

--------------------------------

### CLI Command to Inspect Schema

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/config_schema_contract.md

This command-line interface helper uses Python to inspect the configuration schema. It takes breadcrumbs as a JSON string argument to specify the desired schema node, mirroring the `/schema` API endpoint's functionality.

```bash
python run.py --inspect-schema --schema-breadcrumbs '[{"node":"DesignConfig","field":"graph"}]'
```

--------------------------------

### Configure Multi-Stage Human Review

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/en/nodes/human.md

Chains multiple human review nodes to enforce sequential approval processes.

```yaml
nodes:
  - id: Draft Generator
    type: agent
    config:
      provider: openai
      name: gpt-4o

  - id: Content Review
    type: human
    config:
      description: Please review content accuracy. Enter APPROVED or modification suggestions.

  - id: Final Reviewer
    type: human
    config:
      description: Final confirmation. Enter PUBLISH to publish or REJECT to reject.

edges:
  - from: Draft Generator
    to: Content Review
  - from: Content Review
    to: Final Reviewer
    condition:
      type: keyword
      config:
        any: [APPROVED]
```

--------------------------------

### 注入固定上下文

Source: https://github.com/openbmb/chatdev/blob/main/docs/user_guide/zh/nodes/literal.md

通过 Literal 节点向后续 Agent 节点注入固定的规则指令。

```yaml
nodes:
  - id: Context Injector
    type: literal
    config:
      content: |
        请注意以下规则：
        1. 回答必须简洁明了
        2. 使用中文回复
        3. 如有不确定，请说明
      role: user

  - id: Assistant
    type: agent
    config:
      provider: openai
      name: gpt-4o

edges:
  - from: Context Injector
    to: Assistant
```