User Guide

Concepts and operations for Agents, inference, and related resources under the console Artificial Intelligence menu.

Reading Guide

• For installation and node preparation, see Quick Start.
• For GPU environment, see NVIDIA/CUDA.
• For Agents, see AI Applications: start with OpenClaw / HermesAgent / Dify / ComfyUI, and Application Templates as needed.
• For inference, see AI Inference: start with Quick Start, then Inference Templates / Model Files, and optionally Accelerator Scenarios.

Console Menu and Feature Overview

After entering Artificial Intelligence, main groups include Applications and Inference:

• Applications: Application instances (Dify / OpenClaw / HermesAgent / ComfyUI) and application templates.
• Inference: Inference deployments, inference templates, model files, and inference images for engines such as vLLM / SGLang.
• Images: Container images used when creating templates.

Terminology

• Template: Resource configuration such as CPU/memory/GPU required to run an instance.
• Model Files: Reusable model resources for inference applications.
• Image: Container image for the instance runtime environment, determining functionality and version.

Typical Relationship

Instance = Image + Spec + (optional) Model.

🗃AI Applications

3 items

🗃AI Inference

7 items

🗃AI网关

6 items

🗃容器主机

5 items