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

# Model Mapping Policy & Disaster Recovery

> Configure custom request aliases, provider parameter templates, and automated disaster recovery failover rules.

The **Model Mapping Policy** page (`https://xroute.ai/personal/model-mapping`) is one of XRoute's most powerful enterprise features, allowing developers to define request aliases, enforce consistent parameters, and configure automated failover without changing application client code.

![Model Mapping Policy](https://mintcdn.com/aaa-fac9ff58/ZSwjhSxfbQOFBUl0/images/console/model-mapping.png?fit=max\&auto=format\&n=ZSwjhSxfbQOFBUl0\&q=85\&s=f5975e4eea269595c5da50b4e02b76d8)

XRoute Model Mapping Policy configuration and automated failover routing

## Core Use Cases

* **Client Model Decoupling**: Direct your code to request an abstract name like `my-production-model` or standard `gpt-4o`, while mapping it behind the scenes to any target model (e.g. `Anthropic/claude-fable-5` or `OpenAI/gpt-6-astra`).
* **Zero-Downtime Migration**: Switch underlying foundation models or providers seamlessly in the console without redeploying backend applications.
* **Automated Disaster Recovery (Failover)**: If an upstream provider experiences rate limits (HTTP 429), regional outages (HTTP 502/503), or degraded response times, XRoute immediately retries the request against configured standby backup models.

***

## Policy Parameters

Click **Create Model Alias Mapping** to configure a routing rule:

| Setting | Description |
| :- | :- |
| **New Mapping Name** | The actual model string your applications will submit in the API request (e.g. `company-llm-fast`). |
| **Preferred Target Model** | Primary model and provider to handle normal traffic. |
| **Provider Parameter Template** | Preset parameters (e.g. default `temperature`, `max_tokens`, system prompt overrides). |
| **Model Type** | Modality of the target models (`text`, `image`, `video`). |
| **Disaster Recovery Rules** | Standby model priority list (e.g., fallback from `Anthropic/claude-fable-5` to `Anthropic-SM/claude-fable-5` or `OpenAI/gpt-6-astra` upon error). |
| **Availability & Enable** | Toggle rules active or paused in real time. |
| **Update Time** | Audit timestamp tracking policy revisions. |


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