base on # AI Toolkit for Visual Studio Code  ## What is AI Toolkit AI Toolkit is a powerful extension for Visual Studio Code that streamlines agent development. With AI Toolkit, you can: - 🔍 **Explore and evaluate models** from a wide range of providers—including Anthropic, OpenAI, GitHub—or run models locally using ONNX and Ollama. - ⚡ **Build and test agents in minutes** with prompt generation, quick starters, and seamless MCP tool integrations. Complete features include: | Feature | Description | Screenshot | |---------|-------------|------------| | [Model Catalog](https://code.visualstudio.com/docs/intelligentapps/models) | Browse and access AI models from various sources. Simplified discovery of GitHub, ONNX, Ollama, OpenAI, Anthropic, and Google models. | <img src="https://github.com/user-attachments/assets/e22102a2-562a-4861-8ad2-323040fde3d9" width="350"> | | [Playground](https://code.visualstudio.com/docs/intelligentapps/playground) | Interactive environment for testing AI models. Quick experimentation with model capabilities including multi-modal support. | <img src="https://github.com/user-attachments/assets/83ee9f29-2692-43b0-bbef-ee4c063e79f0" width="350"> | | [Prompt (Agent) Builder](https://aka.ms/AIToolkit/doc/agentbuilder) | Tools for creating and optimizing prompts. Iterative improvement of prompt engineering techniques. | <img src="https://raw.githubusercontent.com/MuyangAmigo/MuyangAmigo/main/assets/aitk_mcp_readme/mcp_debug.gif" width="350"> | | [Bulk Run](https://code.visualstudio.com/docs/intelligentapps/bulkrun) | Execute multiple prompts across selected models. Efficient testing at scale with various inputs. | <img src="https://github.com/user-attachments/assets/1cbc5f5b-6438-4ca6-98de-36f843956baa" width="350"> | | [Evaluate an AI model with a dataset](https://code.visualstudio.com/docs/intelligentapps/evaluation) | Test AI models against datasets using standard metrics. Measure performance with using built-in evaluators such as F1 score, relevance, similarity, and coherence or create your own evaluators. | <img src="https://github.com/user-attachments/assets/e6695e13-25ac-4741-a049-8afcf432e5b4" width="350"> | | [Fine-tune](https://code.visualstudio.com/docs/intelligentapps/finetune) | Customize models for specific use cases. Adapt models to specialized domains and requirements. | <img src="https://github.com/user-attachments/assets/6c1e3c47-c1d9-465a-abf8-3d23dd858d99" width="350"> | ## Getting started  We recommend starting with models hosted by GitHub. - Follow the [installation guide](https://code.visualstudio.com/docs/intelligentapps/overview#_install-and-setup) to set up AI Toolkit for your device. - From the extension tree view, select **CATALOG** > **Models** to explore models available. We recommend to getting started with models hosted by GitHub. - From the model card, select **Try in Playground** to start experimenting the capability of an AI Model. ## Build AI agents The key feature of AI Toolkit is to build AI agents. The agent builder provides a set of tools to help you create and optimize your AI agents. You can use the agent builder to: - ✨ Generate starter prompts with natural language - 🔁 Iterate and refine prompts based on model responses - 🧩 Break down tasks with prompt chaining and structured outputs - ⚡ Test integrations with real-time runs and tool use such as MCP servers - 💡 Generate production-ready code for rapid app development - 🧷 Use variables in prompts - 🧪 Run agents with test cases to validate your agent easily - 📊 Evaluate the accuracy and performance of your agent with built-in or custom metrics - 🔗 Function calling support: Enable agents to invoke external functions dynamically - 🗂️ Agent versioning and version comparison for evaluation results And a lot of features are coming soon, stay tuned for: - 🐞 Local tracing and debugging of agents - 🚀 Deploy your models and agents to Azure AI Foundry - ☁️ Deploy your agent to the cloud Agents can now connect to external tools through MCP (Model Control Protocol) servers, enabling them to perform real-world actions like querying a database, accessing APIs, or executing custom logic. | Feature | Description | Screenshot | |---------|-------------|------------| | Connect to an Existing MCP Server | Use tools from command(stdio) or HTTP (server-sent event) | <img src="https://raw.githubusercontent.com/MuyangAmigo/MuyangAmigo/main/assets/vsc_readme/mcp_existing.gif" width="350"> | | Build and Scaffold a New MCP Server | Start creating your own MCP server from a simple scaffold and test in Agent Builder | <img src="https://raw.githubusercontent.com/MuyangAmigo/MuyangAmigo/main/assets/vsc_readme/scaffold_mcp.gif" width="350"> | ## Feedback and resources We value your feedback to help shape our roadmap. Explore our [developer documentation](https://aka.ms/AIToolkit/doc) for more features, [open issues or share suggestions on GitHub](https://aka.ms/AIToolkit/feedback), or join our [Discord community](https://aka.ms/azureaifoundry/discord) to connect with other developers. AI Toolkit ❤️ Developer Community. ## Data and telemetry The AI Toolkit for Visual Studio Code collects usage data and sends it to Microsoft to help improve our products and services. Read our [privacy statement](https://privacy.microsoft.com/privacystatement) to learn more. This extension respects the `telemetry.enableTelemetry` setting which you can learn more about at [disable telemetry reporting](https://code.visualstudio.com/docs/supporting/faq#_how-to-disable-telemetry-reporting). ", Assign "at most 3 tags" to the expected json: {"id":"12197","tags":[]} "only from the tags list I provide: [{"id":77,"name":"3d"},{"id":89,"name":"agent"},{"id":17,"name":"ai"},{"id":54,"name":"algorithm"},{"id":24,"name":"api"},{"id":44,"name":"authentication"},{"id":3,"name":"aws"},{"id":27,"name":"backend"},{"id":60,"name":"benchmark"},{"id":72,"name":"best-practices"},{"id":39,"name":"bitcoin"},{"id":37,"name":"blockchain"},{"id":1,"name":"blog"},{"id":45,"name":"bundler"},{"id":58,"name":"cache"},{"id":21,"name":"chat"},{"id":49,"name":"cicd"},{"id":4,"name":"cli"},{"id":64,"name":"cloud-native"},{"id":48,"name":"cms"},{"id":61,"name":"compiler"},{"id":68,"name":"containerization"},{"id":92,"name":"crm"},{"id":34,"name":"data"},{"id":47,"name":"database"},{"id":8,"name":"declarative-gui "},{"id":9,"name":"deploy-tool"},{"id":53,"name":"desktop-app"},{"id":6,"name":"dev-exp-lib"},{"id":59,"name":"dev-tool"},{"id":13,"name":"ecommerce"},{"id":26,"name":"editor"},{"id":66,"name":"emulator"},{"id":62,"name":"filesystem"},{"id":80,"name":"finance"},{"id":15,"name":"firmware"},{"id":73,"name":"for-fun"},{"id":2,"name":"framework"},{"id":11,"name":"frontend"},{"id":22,"name":"game"},{"id":81,"name":"game-engine "},{"id":23,"name":"graphql"},{"id":84,"name":"gui"},{"id":91,"name":"http"},{"id":5,"name":"http-client"},{"id":51,"name":"iac"},{"id":30,"name":"ide"},{"id":78,"name":"iot"},{"id":40,"name":"json"},{"id":83,"name":"julian"},{"id":38,"name":"k8s"},{"id":31,"name":"language"},{"id":10,"name":"learning-resource"},{"id":33,"name":"lib"},{"id":41,"name":"linter"},{"id":28,"name":"lms"},{"id":16,"name":"logging"},{"id":76,"name":"low-code"},{"id":90,"name":"message-queue"},{"id":42,"name":"mobile-app"},{"id":18,"name":"monitoring"},{"id":36,"name":"networking"},{"id":7,"name":"node-version"},{"id":55,"name":"nosql"},{"id":57,"name":"observability"},{"id":46,"name":"orm"},{"id":52,"name":"os"},{"id":14,"name":"parser"},{"id":74,"name":"react"},{"id":82,"name":"real-time"},{"id":56,"name":"robot"},{"id":65,"name":"runtime"},{"id":32,"name":"sdk"},{"id":71,"name":"search"},{"id":63,"name":"secrets"},{"id":25,"name":"security"},{"id":85,"name":"server"},{"id":86,"name":"serverless"},{"id":70,"name":"storage"},{"id":75,"name":"system-design"},{"id":79,"name":"terminal"},{"id":29,"name":"testing"},{"id":12,"name":"ui"},{"id":50,"name":"ux"},{"id":88,"name":"video"},{"id":20,"name":"web-app"},{"id":35,"name":"web-server"},{"id":43,"name":"webassembly"},{"id":69,"name":"workflow"},{"id":87,"name":"yaml"}]" returns me the "expected json"