@youdotcom-oss/teams-anthropic

Anthropic SDK integration for Microsoft Teams.ai - Use Claude models (Opus, Sonnet, Haiku) in your Teams.ai applications with just a few lines of code.

Features

Type-Safe Model Selection - Use enums instead of error-prone strings
Streaming Support - Get responses token-by-token with onChunk callback
Function Calling - Auto-execute functions with Claude's tool use
Multi-Part Messages - Send text and handle complex conversations
Full IChatModel Interface - Drop-in replacement for OpenAI models
Configurable - Set temperature, max tokens, and all Anthropic parameters

AI Agent Skills

For Teams Anthropic Integration: Use the teams-anthropic-integration skill to quickly set up Teams apps with You.com MCP server integration.

# Install the Teams Anthropic integration skill
npx skills add youdotcom-oss/agent-skills --skill teams-anthropic-integration

Once installed, ask your AI agent: "Add Anthropic Claude to my Teams app" or "Create a Teams app with You.com MCP and Anthropic"

Supported AI agents: Claude Code, Cursor, Windsurf, Cody, Continue, and more.

See Skill Documentation for complete integration guide.

Getting Started

Get up and running with Claude in your Teams.ai app in 3 quick steps:

1. Install the package

npm install @youdotcom-oss/teams-anthropic @anthropic-ai/sdk

2. Set your API key

Get your API key from console.anthropic.com and set it in your environment:

export ANTHROPIC_API_KEY=your-api-key-here

3. Use it in your app

import { AnthropicChatModel, AnthropicModel } from '@youdotcom-oss/teams-anthropic';

const model = new AnthropicChatModel({
  model: AnthropicModel.CLAUDE_SONNET_4_6,
});

const response = await model.send(
  { role: 'user', content: 'What is the capital of France?' }
);

console.log(response.content); // "The capital of France is Paris."

That's it! Your Teams.ai app now uses Claude models.

MCP Client Integration

Integrate with You.com MCP server for web search and AI capabilities:

import { App } from '@microsoft/teams.apps';
import { ChatPrompt } from '@microsoft/teams.ai';
import { ConsoleLogger } from '@microsoft/teams.common';
import { McpClientPlugin } from '@microsoft/teams.mcpclient';
import { AnthropicChatModel, AnthropicModel } from '@youdotcom-oss/teams-anthropic';

// Validate required environment variables
if (!process.env.YDC_API_KEY) {
  throw new Error('YDC_API_KEY environment variable is required');
}

if (!process.env.ANTHROPIC_API_KEY) {
  throw new Error('ANTHROPIC_API_KEY environment variable is required');
}

const logger = new ConsoleLogger('mcp-client', { level: 'info' });

const prompt = new ChatPrompt(
  {
    instructions: 'You are a helpful assistant with access to web search and AI capabilities.',
    model: new AnthropicChatModel({
      model: AnthropicModel.CLAUDE_SONNET_4_6,
      apiKey: process.env.ANTHROPIC_API_KEY,
    }),
  },
  [new McpClientPlugin({ logger })]
).usePlugin('mcpClient', {
  url: process.env.MCP_SERVER_URL || 'http://localhost:4000/mcp',
  params: {
    headers: {
      'User-Agent': 'teams-ai-mcp-client/1.0.0 (teams.ai; anthropic)',
      Authorization: `Bearer ${process.env.YDC_API_KEY}`,
    },
  },
});

const app = new App();

app.on('message', async ({ send, activity }) => {
  await send({ type: 'typing' });
  const result = await prompt.send(activity.text);
  if (result.content) {
    await send(result.content);
  }
});

app.start().catch(console.error);

Complete template available: node_modules/@youdotcom-oss/teams-anthropic/templates/mcp-client.ts

Usage Examples

Basic Chat

Send a message and get a response:

import { AnthropicChatModel, AnthropicModel } from '@youdotcom-oss/teams-anthropic';

const model = new AnthropicChatModel({
  model: AnthropicModel.CLAUDE_SONNET_4_6,
  requestOptions: {
    max_tokens: 2048,
    temperature: 0.7,
  },
});

const response = await model.send(
  { role: 'user', content: 'Explain quantum computing in simple terms' },
  {
    system: {
      role: 'system',
      content: 'You are a helpful teacher who explains complex topics simply.'
    }
  }
);

console.log(response.content);

Streaming Responses

Get responses token-by-token for a better user experience:

const model = new AnthropicChatModel({
  model: AnthropicModel.CLAUDE_SONNET_4_6,
});

const response = await model.send(
  { role: 'user', content: 'Write a short story about a robot' },
  {
    onChunk: async (delta) => {
      // Stream each token as it arrives
      process.stdout.write(delta);
    },
  }
);

console.log('\n\nFull response:', response.content);

Function Calling

Let Claude call functions to get information:

const model = new AnthropicChatModel({
  model: AnthropicModel.CLAUDE_SONNET_4_6,
});

const response = await model.send(
  { role: 'user', content: 'What is the weather in San Francisco?' },
  {
    functions: {
      get_weather: {
        description: 'Get the current weather for a location',
        parameters: {
          location: { type: 'string', description: 'City name' },
        },
        handler: async (args: { location: string }) => {
          // Your API call here
          return { temperature: 72, conditions: 'Sunny' };
        },
      },
    },
  }
);

console.log(response.content); // Claude uses the function result to answer

Conversation with Memory

Maintain context across multiple messages:

import { LocalMemory } from '@microsoft/teams.ai';

const memory = new LocalMemory();
const model = new AnthropicChatModel({
  model: AnthropicModel.CLAUDE_SONNET_4_6,
});

// First message
await model.send(
  { role: 'user', content: 'My name is Alice' },
  { messages: memory }
);

// Second message - Claude remembers the context
const response = await model.send(
  { role: 'user', content: 'What is my name?' },
  { messages: memory }
);

console.log(response.content); // "Your name is Alice."

Available Models

Choose from the latest Claude models using type-safe enums:

Enum	Model ID	Description
`AnthropicModel.CLAUDE_OPUS_4_6`	`claude-opus-4-6`	Most capable, best for complex tasks and agents
`AnthropicModel.CLAUDE_SONNET_4_6`	`claude-sonnet-4-6`	Best combination of speed and intelligence
`AnthropicModel.CLAUDE_HAIKU_4_5`	`claude-haiku-4-5-20251001`	Fastest, near-frontier intelligence
`AnthropicModel.CLAUDE_OPUS_4_5`	`claude-opus-4-5-20251101`	Previous generation Opus
`AnthropicModel.CLAUDE_SONNET_4_6`	`claude-sonnet-4-5-20250929`	Previous generation Sonnet
`AnthropicModel.CLAUDE_3_HAIKU`	`claude-3-haiku-20240307`	Legacy (deprecated, retiring 2026-04-19)

See all available models with helper functions:

import { getAllModels, getModelDisplayName, getModelFamily } from '@youdotcom-oss/teams-anthropic';

const models = getAllModels();
const displayName = getModelDisplayName(AnthropicModel.CLAUDE_SONNET_4_6); // "Claude Sonnet 4.6"
const family = getModelFamily(AnthropicModel.CLAUDE_HAIKU_4_5); // "haiku"

Configuration Options

Customize the model behavior with configuration options:

AnthropicChatModelOptions

const model = new AnthropicChatModel({
  // Required: Type-safe model selection
  model: AnthropicModel.CLAUDE_SONNET_4_6,

  // Optional: API key (defaults to ANTHROPIC_API_KEY env var)
  apiKey: 'your-api-key',

  // Optional: Custom base URL for proxies
  baseUrl: 'https://your-proxy.com',

  // Optional: Custom headers
  headers: {
    'X-Custom-Header': 'value',
  },

  // Optional: Request timeout in milliseconds
  timeout: 60_000,

  // Optional: Default request options
  requestOptions: {
    max_tokens: 4096,
    temperature: 0.7,
    top_p: 0.9,
    top_k: 40,
  },

  // Optional: Custom logger
  logger: myLogger,
});

Request Options (per message)

const response = await model.send(message, {
  // System message
  system: { role: 'system', content: 'You are a helpful assistant' },

  // Memory for conversation context
  messages: memory,

  // Streaming callback
  onChunk: async (delta) => console.log(delta),

  // Function/tool definitions
  functions: {
    function_name: {
      description: 'Function description',
      parameters: { /* JSON schema */ },
      handler: async (args) => { /* implementation */ },
    },
  },

  // Auto-execute functions (default: true)
  autoFunctionCalling: true,

  // Override default request options
  request: {
    max_tokens: 2048,
    temperature: 0.5,
  },
});

Troubleshooting

Error: "API key is required"

Problem: You're getting an authentication error when trying to use the model.

Solution: Make sure you've set your Anthropic API key:

# Option 1: Environment variable
export ANTHROPIC_API_KEY=your-api-key-here

# Option 2: Pass directly in code
const model = new AnthropicChatModel({
  model: AnthropicModel.CLAUDE_SONNET_4_6,
  apiKey: 'your-api-key-here',
});

Error: "Invalid model identifier"

Problem: You're passing a string instead of using the enum.

Solution: Always use the AnthropicModel enum:

// ✅ Correct
const model = new AnthropicChatModel({
  model: AnthropicModel.CLAUDE_SONNET_4_6,
});

// ❌ Wrong
const model = new AnthropicChatModel({
  model: 'claude-sonnet-4-6', // Type error!
});

Streaming not working

Problem: You're not seeing token-by-token responses.

Solution: Make sure you provide the onChunk callback:

const response = await model.send(message, {
  onChunk: async (delta) => {
    // This callback is required for streaming
    process.stdout.write(delta);
  },
});

Functions not executing automatically

Problem: Function calls are returned but not executed.

Solution: Functions auto-execute by default. If you want to control execution manually, set autoFunctionCalling: false:

const response = await model.send(message, {
  functions: myFunctions,
  autoFunctionCalling: false, // Disable auto-execution
});

// Now response.function_calls will contain the calls to execute manually

API Reference

API documentation is provided via TypeScript types and TSDoc comments in the source code. See the examples above and TypeScript intellisense in your IDE for complete API details.

Templates

See the templates directory for integration templates:

mcp-client.ts - Complete MCP client integration with custom user agent headers

Access the template after installation:

# Template location in your node_modules
node_modules/@youdotcom-oss/teams-anthropic/templates/mcp-client.ts

Contributing

See AGENTS.md for development setup and contribution guidelines.

License

MIT

Support

Issues: GitHub Issues
Email: support@you.com

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

@youdotcom-oss/teams-anthropic

Features

AI Agent Skills

Getting Started

1. Install the package

2. Set your API key

3. Use it in your app

MCP Client Integration

Usage Examples

Basic Chat

Streaming Responses

Function Calling

Conversation with Memory

Available Models

Configuration Options

Troubleshooting

Error: "API key is required"

Error: "Invalid model identifier"

Streaming not working

Functions not executing automatically

API Reference

Templates

Contributing

License

Support

FilesExpand file tree

README.md

Latest commit

History

README.md

File metadata and controls

@youdotcom-oss/teams-anthropic

Features

AI Agent Skills

Getting Started

1. Install the package

2. Set your API key

3. Use it in your app

MCP Client Integration

Usage Examples

Basic Chat

Streaming Responses

Function Calling

Conversation with Memory

Available Models

Configuration Options

Troubleshooting

Error: "API key is required"

Error: "Invalid model identifier"

Streaming not working

Functions not executing automatically

API Reference

Templates

Contributing

License

Support