Application types

更新时间:
复制 MD 格式

Large language models have limitations in processing private knowledge, accessing real-time information, following fixed procedures, and planning complex tasks. To overcome these challenges, Model Studio provides three application types: agent applications, workflow applications, and high-code applications. By integrating capabilities such as knowledge base retrieval, external tool calls, and memory, Model Studio lets you build AI applications that solve business problems.

Quick selection

  • Agent application:

    • Driven by prompts, these applications autonomously understand user intent, make decisions, and call external tools such as knowledge bases and MCP services to complete tasks.

    • Best for open-ended conversational applications, such as customer support, knowledge Q&A, task assistance, and travel planning.

  • Workflow application:

    • These applications use visual node orchestration to chain multi-step, complex tasks into a stable, controllable, and repeatable execution path.

    • Suitable for automating fixed processes, such as generating reports, processing orders, managing multi-step approval flows, and labeling data.

  • High-code application:

    • Designed for professional developers, this mode lets you build and deploy AI backend services using Python code. You can deploy to the cloud with one click and automatically integrate operational capabilities.

    • Best for scenarios that require customization, such as deploying proprietary algorithms, integrating complex systems, and building AI backend services.

Application comparison

Dimension

Agent

Workflow

High-code

Development approach

Natural language configuration (no-code)

Visual node orchestration (low-code)

Python coding (pro-code)

Key characteristics

Autonomous AI decision-making and dynamic planning
The LLM autonomously plans task steps based on the prompt.

Precisely controlled by a predefined process
Each step is defined by a preset node, ensuring deterministic logic.

Fully controlled by code
All logic and execution paths are defined in the code.

Best for

Business users, product managers, and operations personnel

IT operations personnel, business analysts, and implementation consultants

AI engineers and developers

Skill requirements

Low

Medium

High

Get started

You can create an application in the console, explore practical examples, or integrate the application into your existing system using APIs: