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

# Platform integrations

These integrations make Sprites available inside another platform rather than through a standalone SDK in your application.

## Anthropic Connectors Directory

The [Sprites listing in Anthropic's Connectors Directory](https://claude.ai/directory/sprites) connects Claude directly to the hosted Sprites MCP server.

Before starting, sign in to Claude with access to connectors and have a Fly.io account with access to the Sprites organization you want to use.

1. Open the Sprites listing in the directory.
2. Select **Connect to Claude**.
3. Complete the Fly.io browser OAuth flow.
4. Choose the intended Fly.io organization and review the connector's name prefix, creation limit, and access level.
5. Ask Claude to list your Sprites.

An empty list still confirms that the connector authenticated successfully. You do not need to install the Sprites CLI or paste an API token into Claude.

If you cannot use the directory listing, add `https://sprites.dev/mcp` as a custom remote MCP connector instead. See [Remote MCP Server](/sprites/integrations/remote-mcp/) for the permission model.

<Info>
  The Anthropic Connectors Directory integration is separate from [Claude Managed Agents](/sprites/integrations/claude-managed-agents/). The directory lets Claude call the hosted Sprites MCP tools; Managed Agents run a self-hosted per-session tool environment inside a Sprite.
</Info>

## QM by Y Combinator

[QM](https://github.com/yc-software/qm) is an open-source agent platform with a Sprites sandbox backend. QM's control plane and UI stay in the deployment you manage; agent commands run in Sprites when the backend is selected. Its published CLI requires Node.js 24 or newer.

Before starting, have access to your chosen deployment infrastructure, the model credentials required by your QM configuration, and a [Sprites API token](https://sprites.dev/account). See [QM's deployment guide](https://github.com/yc-software/qm/blob/main/deployment.md) for the full prerequisites.

Start with QM's deployment initializer. Replace `<slug>` with your organization slug and `<fly-or-aws>` with `fly` or `aws`:

```sh theme={null}
npm exec --yes --package=@yc-software/qm@latest -- \
  qm init . --org <slug> --target <fly-or-aws>
npm install
```

In the generated deployment configuration, select the Sprites backend and name the Fly app in which QM's agents execute:

```json theme={null}
{
  "sandbox": {
    "backend": "sprites",
    "app": "your-sandbox-app"
  }
}
```

Provide `SPRITES_TOKEN` through QM's deployment secret workflow. The initializer walks through infrastructure, credentials, deployment, and live verification for the selected target.

After deployment, follow the [QM CLI's live checks](https://github.com/yc-software/qm/blob/main/cli/README.md), then ask QM to run `pwd` in its sandbox. Verify a successful sandbox tool result; loading QM's UI alone does not confirm that the Sprites backend works.

<Warning>
  QM is a complete platform deployment, not a one-click connector. Follow QM's current deployment guide and run its generated checks before treating a deployment as production-ready.
</Warning>