Server deployments in Metorial define which integrations your AI agents can access. This guide explains how to set up and use server deployments.

What Are Server Deployments?

Server deployments are collections of MCP servers (integrations) that you configure in the Metorial dashboard. Each deployment can include specific integrations like GitHub, Slack, Google Drive, and more.

Creating a Server Deployment

1. Access the Dashboard

Log into your Metorial account and navigate to the "Server Deployments" section.

2. Create a New Deployment

Click "New Deployment" and give it a descriptive name (e.g., "Production Integrations" or "GitHub Tools").

3. Add Integrations

Browse the 600+ available integrations and add the ones your agent needs. You can always add or remove integrations later.

4. Copy the Deployment ID

Once created, copy your deployment ID. You'll need this in your code.

Using Server Deployments in Code

Pass your deployment ID when creating a provider session:

await metorial.withProviderSession(
  metorialAiSdk,
  {
    serverDeployments: ['your-deployment-id-here']
  },
  async session => {
    // Your agent code here
  }
);

Using Multiple Deployments

You can use multiple deployments to give your agent access to different sets of tools:

{
  serverDeployments: [
    'github-tools-deployment-id',
    'communication-tools-deployment-id'
  ]
}

Best Practices

• Create separate deployments for different environments (dev, staging, production)
• Group related integrations together
• Only include integrations your agent actually needs
• Test new integrations in a development deployment first