Maze

Contents:

  • Overview
    • What is Maze?
    • Core Features
      • Task-level
      • Resource Management
      • Distributed Deployment
  • Installation
    • Prerequisites
    • Using a Virtual Environment (Recommended)
    • Install from Source
    • Verify Installation
  • Quick Start
    • Your First Workflow
      • Step 1: Define Your Tasks
      • Step 2: Create a Workflow
      • Step 3: Add Tasks to the Workflow
      • Step 4: Run the Workflow
    • Complete Example
      • Expected Output
    • Next Steps
    • Key Concepts to Remember
  • Client Guides
    • MaClient
      • Core Concepts
        • Workflow
        • Task
        • Edge
      • Workflow Patterns
        • Single End Task Workflow
        • Multiple End Tasks Workflow
        • Diamond Pattern Workflow
        • Complex Workflows
      • Working with Task Outputs
        • Referencing Outputs
        • Multiple Outputs
      • Resource Configuration
        • Configuring Task Resources
      • Handling Workflow Results
        • Run ID Concept
        • Simple Result Display
        • Real-time Result Streaming (Advanced)
        • Collecting Results
        • Query Specific Task Results
        • Results Caching
      • Best Practices
        • Task Design
        • Workflow Design
        • Performance Optimization
        • Debugging Tips
      • Next Steps
    • LanggraphClient
      • Initialization
      • Defining Tasks
      • Building the Graph
      • Executing the Workflow
    • LLM Instance Management
      • start_llm_instance
      • query_llm_instance
      • stop_llm_instance
      • Complete Example
  • Examples
  • API References
    • MaClient API Reference
      • MaClient
        • MaClient
      • MaWorkflow
        • MaWorkflow
      • MaTask
        • MaTask
      • TaskOutput
        • TaskOutput
      • TaskOutputs
        • TaskOutputs
      • Decorators
        • @task
      • Data Types
        • Input Schema Types
        • Resource Configuration
      • See Also
    • LanggraphClient API Reference
  • Maze Playground
    • Overview
      • Key Features
      • Architecture
    • Getting Started
      • Starting Playground
      • Starting Without Playground
      • Stopping Services
    • Using the Playground
      • Interface Overview
      • Creating Your First Workflow
        • Step 1: Add Task Nodes
        • Step 2: Connect Tasks
        • Step 3: Configure Task Inputs
        • Step 4: Execute Workflow
        • Expected Results
      • Built-in Tasks
        • task1
        • task2
      • Advanced Workflow Patterns
        • Sequential Workflow
        • Parallel Workflow
        • Complex Workflow
    • Backend Implementation
      • Bridge Architecture
      • Key Components
        • Client Module
        • Task Metadata
        • WebSocket Communication
        • Process Management
      • Design Decisions
    • Troubleshooting
      • Common Issues
    • Next Steps
    • Additional Resources
  • Sandbox
    • Starting the Sandbox Service
    • Basic Usage
    • Resource Allocation
    • Example: Running Complex Code
    • Concurrent Usage
Maze
  • Client Guides
  • View page source

Client Guides

Maze provides multiple client options to build and orchestrate AI workflows. Choose the client that best fits your use case.

  • MaClient
    • Core Concepts
    • Workflow Patterns
    • Working with Task Outputs
    • Resource Configuration
    • Handling Workflow Results
    • Best Practices
    • Next Steps
  • LanggraphClient
    • Initialization
    • Defining Tasks
    • Building the Graph
    • Executing the Workflow
  • LLM Instance Management
    • start_llm_instance
    • query_llm_instance
    • stop_llm_instance
    • Complete Example
Previous Next

© Copyright 2025, HUST.

Built with Sphinx using a theme provided by Read the Docs.