Vercel AI SDK

Build an agent using the Vercel AI SDK that reads a user’s Gmail inbox. Use tool() and jsonSchema() from the ai package to wrap Scalekit tools. No manual schema conversion needed.

Install

npm install @scalekit-sdk/node ai @ai-sdk/openai

Initialize

import { ScalekitClient } from '@scalekit-sdk/node';
import { ConnectorStatus } from '@scalekit-sdk/node/lib/pkg/grpc/scalekit/v1/connected_accounts/connected_accounts_pb';

const scalekit = new ScalekitClient(
  process.env.SCALEKIT_ENV_URL!,
  process.env.SCALEKIT_CLIENT_ID!,
  process.env.SCALEKIT_CLIENT_SECRET!,
);

Connect the user to Gmail

const { connectedAccount } = await scalekit.actions.getOrCreateConnectedAccount({
  connectionName: 'gmail',
  identifier: 'user_123',
});
if (connectedAccount?.status !== ConnectorStatus.ACTIVE) {
  const { link } = await scalekit.actions.getAuthorizationLink({ connectionName: 'gmail', identifier: 'user_123' });
  console.log('Authorize Gmail:', link);
}

See Authorize a user for production auth handling.

Run the agent

import { generateText, jsonSchema, stepCountIs, tool } from 'ai';
import { openai } from '@ai-sdk/openai';

const { tools: scopedTools } = await scalekit.tools.listScopedTools('user_123', {
  filter: { connectionNames: ['gmail'] },
});

const tools = Object.fromEntries(
  scopedTools.map(t => [
    t.tool.definition.name,
    tool({
      description: t.tool.definition.description,
      parameters: jsonSchema(t.tool.definition.input_schema ?? { type: 'object', properties: {} }),
      execute: async (args) => {
        const result = await scalekit.actions.executeTool({
          toolName: t.tool.definition.name,
          identifier: 'user_123',
          toolInput: args,
        });
        return result.data;
      },
    }),
  ]),
);

const { text } = await generateText({
  model: openai('gpt-4o'),
  tools,
  stopWhen: stepCountIs(5),
  prompt: 'Fetch my last 5 unread emails and summarize them',
});
console.log(text);

Use MCP instead

The Vercel AI SDK supports MCP via experimental_createMCPClient. Pass a Scalekit-generated MCP URL to connect without any tool schema setup:

import { experimental_createMCPClient, generateText } from 'ai';
import { openai } from '@ai-sdk/openai';

const mcpClient = await experimental_createMCPClient({
  transport: {
    type: 'sse',
    url: mcpUrl,  // from actions.mcp.ensure_instance()
  },
});

const tools = await mcpClient.tools();

const { text } = await generateText({
  model: openai('gpt-4o'),
  tools,
  stopWhen: stepCountIs(5),
  prompt: 'Fetch my last 5 unread emails and summarize them',
});
await mcpClient.close();
console.log(text);

See Generate user MCP URLs to get mcpUrl.