What is AI Library MCP?

AI Library MCP connects your coding agent to AI Library's managed tool infrastructure through the Model Context Protocol.

Instead of wiring together search, document parsing, transcription, storage, vector search, and agent orchestration yourself, you install one MCP server and get a production-ready toolset inside your editor.

Quick Start

1. Get your API key from mcp.ailibrary.ai.
2. Install AI Library MCP from the VS Code Marketplace.
3. Open GitHub Copilot Chat and use any AI Library tool.
4. Enter your API key when VS Code prompts you.

VS Code + GitHub Copilot
    |
    v
  AI Library MCP
    |
    v
Search | Documents | Notes | Knowledge | Speech | Agents

Why use AI Library?

| Without AI Library | With AI Library | |------------------|-----------------------------------------------------------------| | Build and maintain your own app database stack | Managed production-grade storage | | Stand up and tune your own vector database | Managed vector search and retrieval | | Design RAG pipelines from scratch | Production-informed retrieval patterns | | Stitch together document parsing services | Built-in document parsing optimized for LLMs and lower token usage | | Integrate separate transcription providers | Built-in transcription | | Assemble and tune web search workflows | Built-in web search | | Manage tool orchestration yourself | Unified MCP tools for coding agents |

Where can I use it?

AI Library MCP works with MCP-compatible coding agents.

• GitHub Copilot in VS Code
• Claude Code

Install

1. Go to mcp.ailibrary.ai and create your API key.
2. Fund your account. New signups receive credits to get started.
3. Install AI Library MCP in your coding agent of choice.

For GitHub Copilot in VS Code:

• Install from the marketplace: https://marketplace.visualstudio.com/items?itemName=AILibrary.ai-library-mcp
• On first use, the extension prompts you for your API key automatically.

For Claude Code:

claude mcp add ailibrary-mcp -e AI_LIBRARY_KEY=your_key_here -- npx -y @ailibrary/mcp

🧰 Key Skills

AI Library MCP exposes a focused set of managed tools for common coding-agent workflows. These tools are available through your MCP client and are backed by AI Library services.

Web and Documents

| Tool | What it does | |------|--------------| | websearch | Search the web or images with optional filters such as domain, country, and time range | | read-documents | Extract LLM-friendly text from documents and URLs | | upload-files | Upload local files and return hosted URLs for downstream processing |

Speech and Audio

| Tool | What it does | |------|--------------| | transcribe | Convert audio files to text | | generate-speech | Convert text into generated speech |

Agents and Structured Workflows

| Tool | What it does | |------|--------------| | create-agent | Create a managed AI agent with instructions and optional form schema | | get-agent | Retrieve an existing agent by namespace | | update-agent | Update an agent's instructions, model, knowledge, or form wiring | | chat-agent | Chat with an existing agent and optionally receive structured output | | update-form | Update a form schema used for structured responses |

Knowledge Base

| Tool | What it does | |------|--------------| | agent-knowledge-add | Add files or content to an agent knowledge base | | agent-knowledge-list | List knowledge sources attached to an agent | | agent-knowledge-search | Search across an agent's indexed knowledge | | delete-knowledge-sources | Remove knowledge sources from an agent |

Database

| Tool | What it does | |------|--------------| | notes-create | Create a record attached to a resource | | notes-get | Fetch a record by ID | | notes-list | List records for a resource | | notes-update | Update an existing record | | notes-delete | Delete a record |

File Storage

| Tool | What it does | |------|--------------| | upload-files | Upload a file | | list-files | List uploaded files | | delete-file | Delete a file by ID |

These tools let your coding agent search the web, parse documents, manage knowledge, store state, and interact with structured AI workflows without custom backend plumbing.

Why teams use it

• Managed backend instead of custom glue code
• Built-in tools for search, retrieval, parsing, notes, and audio
• Structured agent workflows with forms and knowledge bases
• Faster iteration inside the editor with fewer external integrations

Notes

• This extension requires your AI Library API key from mcp.ailibrary.ai.
• On first use, VS Code automatically prompts for your key via a password input box.
• To set or rotate the key at any time: Ctrl/Cmd+Shift+P → AI Library MCP: Set API Key.
• To remove the key: AI Library MCP: Clear API Key.
• Keys are stored in VS Code Secret Storage and never written to settings.json.
• Skills are bundled from ../ai-library-plugin/skills at build time.