Visual Studio Code (VS Code) has just released a Preview version of its AI Toolkit, introducing a game-changing feature: MCP (Model Context Protocol) Server integration. This update is designed to empower developers, data scientists, and AI engineers to build more intelligent, connected AI applications seamlessly within the VS Code environment. Whether you’re just starting with generative AI or are an experienced developer, this toolkit has something to offer for everyone.
What is the AI Toolkit for VS Code?
The AI Toolkit is a powerful extension that simplifies the process of building, testing, and deploying AI-powered solutions. It supports integration with major AI providers like OpenAI, Anthropic, and Google, while also allowing local model usage with ONNX and Ollama. With features like model fine-tuning, prompt engineering, and evaluation, it’s a one-stop shop for AI development.
The latest Preview version takes the functionality even further by introducing MCP Server support, enabling developers to connect their AI agents to external APIs and services, making them truly capable of performing actions beyond text generation.
Key Highlights of the Latest Release
1. MCP Server Integration
The MCP Server is a standout feature in this release. It allows developers to:
- Connect to external APIs and services: For example, you can integrate your agent with a database, web service, or any external application.
- Build custom MCP servers: Developers can use Python or TypeScript scaffolds provided by the toolkit to create their own servers for specific needs.
- Enhance agent capabilities: With MCP servers, agents can perform tasks like fetching live weather data, searching files, or even interacting with cloud services.
2. Agent Builder with MCP Support
The Agent Builder (formerly Prompt Builder) now supports MCP integrations. This feature enables users to:
- Create sophisticated prompts with structured outputs.
- Chain prompts for complex workflows.
- Integrate MCP tools for dynamic agent functionality.
- Quickly test and debug prompts with real-time feedback.
3. Improved Playground for Real-Time Testing
The AI Toolkit’s Playground has been upgraded to offer:
- Side-by-side model comparisons, enabling developers to evaluate responses from different AI models in real-time.
- Multi-modal input support, including text, images, and file attachments.
- Configurable inference parameters, such as temperature and token limits, for fine-tuning model behavior.
4. Fine-Tuning and Model Conversion
Developers can now fine-tune models locally or in the cloud using Azure Container Apps for GPU support. The toolkit also provides tools to:
- Convert models from platforms like Hugging Face into ONNX for efficient local deployment.
- Optimize models for CPUs, GPUs, or NPUs.
Who Should Use the AI Toolkit?
This update caters to a broad audience, including:
- Developers: Build AI-powered features in apps, websites, or mobile prototypes.
- AI Engineers: Fine-tune and deploy models with domain-specific data.
- Data Scientists: Evaluate models, compare performance, and optimize workflows.
- Educators and Students: Learn and teach AI concepts with hands-on tools.
Getting Started with MCP Servers
To start using MCP servers in the AI Toolkit:
- Install the AI Toolkit from the Visual Studio Marketplace. After installation, the AI Toolkit icon will appear in the Activity Bar.
- In the Tools section, click + MCP Server to add a server.
- Choose from:
- Featured MCP Servers: Pre-built servers for common tasks.
- Connect Existing MCP Server: Link to an already running server.
- Build a New MCP Server: Use the provided scaffold to create your own in Python or TypeScript.
Once configured, the MCP server can be used in the Agent Builder to extend your agent’s functionality.
Why MCP Server Support Matters
MCP servers unlock new possibilities for AI applications by enabling real-world interactivity. For example:
- Dynamic Data Retrieval: An agent can fetch the latest stock prices or weather data.
- File Management: Automate tasks like searching and organizing local files.
- Custom APIs: Seamlessly integrate third-party services like payment gateways or CRMs.
This feature transforms static AI agents into dynamic, multi-functional tools.
How to Try the Preview Version
Getting started is easy:
- Install the AI Toolkit for VS Code from the Visual Studio Marketplace.
- Explore the Playground to test models and parameters.
- Use the Agent Builder to create agents and integrate MCP servers.
- Fine-tune models locally or in the cloud to customize them for your use case.
For detailed instructions, visit the AI Toolkit documentation.
Final Thoughts
The AI Toolkit for VS Code with MCP Server integration is a significant step forward in making AI application development more accessible and powerful. Whether you’re building a simple chatbot or a full-fledged intelligent system, this toolkit provides all the tools you need in one place.
This Preview version is a must-try for anyone serious about leveraging generative AI in their projects. Don’t wait—download the AI Toolkit today and take your AI development to the next level!
Johngai’s Tech Forge 