Real-world examples of using Metorial with AI SDK to build powerful applications.

Personal Assistant

Build an AI assistant that manages calendars, emails, and tasks:

const metorial = new Metorial({
  apiKey: process.env.METORIAL_API_KEY
});

await metorial.withProviderSession(
  metorialAISDK,
  { 
    serverDeployments: ['calendar-email-tasks-deployment'] 
  },
  async (session) => {
    const result = await generateText({
      model: openai('gpt-4'),
      messages: [
        { 
          role: 'system', 
          content: 'You are a personal assistant with access to calendar, email, and task management.'
        },
        { 
          role: 'user', 
          content: 'What do I have scheduled this week and are there any urgent emails?'
        }
      ],
      tools: session.tools,
      maxSteps: 10
    });
    
    console.log(result.text);
  }
);

Integrations used: Google Calendar, Gmail, Todoist

Customer Support Bot

Create a support bot that can access customer data and create tickets:

await metorial.withProviderSession(
  metorialAISDK,
  { 
    serverDeployments: ['crm-support-deployment'] 
  },
  async (session) => {
    const result = await streamText({
      model: anthropic('claude-3-5-sonnet-20241022'),
      messages: [
        { 
          role: 'system', 
          content: 'You are a customer support agent. Help customers and create support tickets when needed.'
        },
        { 
          role: 'user', 
          content: customerMessage
        }
      ],
      tools: session.tools
    });
    
    // Stream response to customer
    for await (const chunk of result.textStream) {
      process.stdout.write(chunk);
    }
  }
);

Integrations used: Salesforce, Zendesk, Intercom

Code Review Assistant

Build a bot that reviews code and creates issues:

await metorial.withProviderSession(
  metorialAISDK,
  { 
    serverDeployments: ['dev-tools-deployment'] 
  },
  async (session) => {
    const result = await generateText({
      model: openai('gpt-4'),
      messages: [
        { 
          role: 'user', 
          content: `Review the recent pull request #${prNumber} and create issues for any problems found.`
        }
      ],
      tools: session.tools,
      maxSteps: 15
    });
    
    console.log(result.text);
  }
);

Integrations used: GitHub, Jira, Linear

Meeting Summarizer

Automatically summarize meetings and share notes:

await metorial.withProviderSession(
  metorialAISDK,
  { 
    serverDeployments: ['productivity-deployment'] 
  },
  async (session) => {
    const result = await generateText({
      model: openai('gpt-4'),
      messages: [
        { 
          role: 'user', 
          content: 'Summarize today\'s team meeting and share the notes in our Slack channel.'
        }
      ],
      tools: session.tools,
      maxSteps: 8
    });
    
    console.log(result.text);
  }
);

Integrations used: Google Calendar, Notion, Slack

Data Analysis Agent

Create an agent that can query databases and create visualizations:

await metorial.withProviderSession(
  metorialAISDK,
  { 
    serverDeployments: ['data-analytics-deployment'] 
  },
  async (session) => {
    const result = await generateText({
      model: anthropic('claude-3-5-sonnet-20241022'),
      messages: [
        { 
          role: 'user', 
          content: 'Analyze last month\'s sales data and create a summary report in Google Sheets.'
        }
      ],
      tools: session.tools,
      maxSteps: 12
    });
    
    console.log(result.text);
  }
);

Integrations used: PostgreSQL, Google Sheets, Airtable

Project Manager Bot

Manage projects across multiple tools:

await metorial.withProviderSession(
  metorialAISDK,
  { 
    serverDeployments: ['project-management-deployment'] 
  },
  async (session) => {
    const result = await generateText({
      model: openai('gpt-4o'),
      messages: [
        { 
          role: 'system', 
          content: 'You are a project manager assistant. Help coordinate tasks and keep projects on track.'
        },
        { 
          role: 'user', 
          content: 'What\'s the status of the Q4 product launch? Any blockers?'
        }
      ],
      tools: session.tools,
      maxSteps: 10
    });
    
    console.log(result.text);
  }
);

Integrations used: Jira, Asana, Linear, GitHub, Slack

Multi-Step Workflows

Handle complex workflows with multiple tool calls:

await metorial.withProviderSession(
  metorialAISDK,
  { 
    serverDeployments: ['full-integration-deployment'] 
  },
  async (session) => {
    const result = await generateText({
      model: openai('gpt-4'),
      messages: [
        { 
          role: 'user', 
          content: `
            1. Check my calendar for tomorrow
            2. Find a 30-minute slot for a meeting with the design team
            3. Create the calendar event
            4. Send a Slack message to #design-team about the meeting
            5. Create a Notion page with the meeting agenda
          `
        }
      ],
      tools: session.tools,
      maxSteps: 20  // Allow enough steps for complex workflow
    });
    
    console.log(result.text);
  }
);

Integrations used: Google Calendar, Slack, Notion

Key Takeaways

• Metorial works with any AI SDK-compatible model
• Combine multiple integrations in a single deployment
• Use maxSteps to control workflow complexity
• System messages help guide agent behavior
• Stream responses for better user experience
• One deployment can power multiple use cases

Building Your Own

To create your own use case:

1. Identify which integrations you need
2. Create a deployment with those integrations
3. Configure authentication for each integration
4. Write clear system prompts to guide behavior
5. Test with various user inputs
6. Adjust maxSteps based on complexity
7. Deploy to production

With Metorial handling integrations, you can focus on creating unique AI experiences for your users.