MCPSDKApiClient

MCPSDKApiClient is the main entry point for interacting with the MCPSDK.ai API.

1. Installation

npm install toolsdk

2. Initialization

Initialize the client with your API key.

import { MCPSDKApiClient } from 'toolsdk/api';

const client = new MCPSDKApiClient({
  apiKey: 'your-api-key', // process.env.TOOLSDK_AI_API_KEY
});

3. Methods

3.1 package(packageKey, version?, envs?)

Access a specific package (MCP Server) to run tools or get tool configurations.

• packageKey: The unique identifier of the package (e.g., github, twilio).
• version: (Optional) Specific version of the package.
• envs: (Optional) Environment variables to pass to the package.

Returns a Package instance.

const github = client.package('github');

3.2 Package.run(options)

Execute a tool from the package.

const result = await client.package('github').run({
  toolKey: 'create-issue',
  inputData: {
    title: 'New Issue',
    body: 'Issue description'
  }
});

3.3 Package.getAISDKTool(toolKey)

Get a tool compatible with Vercel AI SDK (ai package).

const tool = await client.package('github').getAISDKTool('create-issue');

3.4 Package.getAISDKToolSet()

Get all tools from the package as a Vercel AI SDK compatible toolset.

const tools = await client.package('github').getAISDKToolSet();

3.5 Package.getOpenAISDKTools()

Get all tools from the package in OpenAI function calling format.

const tools = await client.package('github').getOpenAISDKTools();

3.6 packages.pages(params)

List available packages in the registry.

const packages = await client.packages.pages({ pageNo: 1, pageSize: 20 });

3.7 packageInstance(instanceId)

Access a specific package instance by ID.

const instance = client.packageInstance('instance-id');