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

# Manifest documentation

> Open-source LLM router for agents. Sends each query to the cheapest model that can handle it, often saving up to 70% on inference costs.

Manifest is a smart model router for agents and AI applications that redirects each query to the right model, saving up to 70% in AI costs.

This documentation will help you to get started and use Manifest to use AI efficiently and reduce your inference costs.

## Key features

<CardGroup cols={2}>
  <Card title="Routing" icon="split" href="/routing">
    Routing based on complexity, specificity and custom HTTP headers
  </Card>

  <Card title="Fallbacks" icon="life-buoy" href="/fallback">
    Fallback to different models on failure.
  </Card>

  <Card title="Set limits" icon="shield-alert" href="/set-limits">
    Get email alerts or block requests when spending crosses a threshold.
  </Card>

  <Card title="All your providers" icon="layers" href="/providers/api-key-providers">
    Plug in every kind of provider: API key, subscription-based, local models or
    custom.
  </Card>
</CardGroup>

## Cloud or self-hosted?

Manifest is an open source project that comes in 2 ways: cloud and self-hosted. We recommend the [cloud version](https://app.manifest.build) for newcomers and the [self-hosted version](./self-hosted) for advanced users.

The key difference is that the cloud version runs on our servers, making it easy to setup whereas the self-hosted version must be installed in your local machine of infrastructure.