> ## Documentation Index
> Fetch the complete documentation index at: https://docs.scitix.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Manage LLM Inference Templates

> Create, reuse, and manage LLM Inference templates in Siverse to retain Models, Engines, probes, monitoring, and Advanced Config.

LLM inference templates are used to create reusable service configurations. After creating a template, you can enable **Use Recommended Template** when creating an LLM inference service to quickly reuse the Model, Engine, Execute Type, probe, Monitor, and some Advanced Config settings.

## Create Template

### 1. Go to the creation page

1. Go to **Inference > LLM Inference Template**.
2. Select a **Region** and **Cluster**.
3. Click **Create**.

### 2. Enter configuration information

#### Basic Setting

| Configuration | Description |
| - | - |
| **Template Name** | Used to identify and select the template when creating a service. It is recommended to include the model, engine, or deployment type. |
| **Is Public** | Controls whether the template can be viewed or reused by other users. |

#### Model and Engine

| Configuration | Description |
| - | - |
| **Model** | Select the model source, model version, and access path. |
| **LoRA Model** | To load a LoRA Adapter, enable it and configure the corresponding model. |
| **Engine** | Select a configured inference engine and version. To add or maintain engine versions, see [Manage LLM Engine Config](/siverse/inference/manage-llm-inference-engine-configurations.md). |
| **Engine Params** | Add startup parameters for the inference engine. System parameters required for distributed deployment and PD disaggregation are typically generated by the platform and do not need to be entered again. |
| **Served Model Name** | Set the model name exposed by the service. |

#### Probes, Monitor, and Advanced Config

| Configuration | Description |
| - | - |
| **Liveness Probe** | Determines whether the service process is alive. If it fails, the platform can restart the instance according to the policy. |
| **Readiness Probe** | Determines whether the service can receive traffic. If it fails, the platform no longer assigns new requests to the instance. |
| **Metrics** | Configures the metrics collection path and port for monitoring, scaling, or alerts. |
| **Environment Variables** | Injects runtime configuration for the service. |
| **Storage** | Mounts additional volumes. |
| **Service Port** | Configures the port on which the service listens or is exposed. |
| **Auto Scaling Configuration** | Configures service replica scaling policies based on time or metrics. |
| **Gateway Configuration** | Configures the service access entry point, authentication, and routing. |
| **Whether to Send Alerts** | Controls whether service exceptions enter the alerting workflow. |

Probes support check methods such as **HTTP**, **TCP**, and **Exec**. If the service provides a health check endpoint, configure a **Readiness Probe** first to prevent traffic from being routed to the service before the Model has finished loading. If automatic recovery is also required when the process hangs or encounters an error, configure a **Liveness Probe**.

### 3. Submit Creation

Click **Submit**. After it is created, you can select this template and reuse its configuration when creating an LLM inference service.

## Create from a Service Template

If an existing service configuration is suitable for reuse, you can save it as a template from the **Create as Template** entry of the target LLM inference service. Before saving, verify that the model path, engine, probes, monitoring, gateway, and public access scope are suitable for reuse by subsequent services.

## Manage Templates

After locating the target template under **Inference > LLM Inference Templates**, you can edit, clone, or delete it as needed.

| Actions | Description |
| - | - |
| **Edit** | Modify the template configuration. After making changes, recheck the Model, Engine, and Advanced Config before next using the template to create a service. |
| **Clone** | Copy an existing template to create a new one. This is suitable when you want to retain the original template while adjusting the Model, Engine, or Advanced Config. |
| **Delete** | Delete templates that are no longer in use. Before deletion, confirm that your team no longer relies on the template to create services. |

## Next Steps

* [Create an LLM inference service](/siverse/inference/create-llm-inference-services.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.