The Yonote MCP server acts as a bridge between AI assistants and the Yonote workspace, a collaborative platform often used as an alternative to Notion. This tool allows an AI model to navigate through a user’s organized notes and folders, making it possible for the assistant to find, read, and summarize documents on demand. By connecting these two systems, users can turn their personal or team documentation into a searchable knowledge base that an AI can interact with directly. Moving into more specific capabilities, the server provides three primary tools: `documents_list`, `documents_info`, and `collections_list`. These functions enable an LLM to programmatically browse through various document collections, filter results with limits and offsets, and retrieve the full content of specific files using their unique IDs. This structured access allows an AI to handle complex tasks, such as cross-referencing information between different documents or keeping track of project updates stored within the Yonote ecosystem. For developers and power users, this MCP server is built on the FastMCP framework and requires Python 3.13 or higher. It is designed for easy deployment, supporting both manual installation via the `uv` package manager and automated setup through Smithery. Because it adheres to the Model Context Protocol, the server can be integrated into environments like Claude Desktop or Cursor, providing AI agents with the necessary permissions and API tools to securely fetch real-time data from the Yonote API. This makes it an essential utility for anyone looking to build a more context-aware AI workflow using their existing documentation.
1. Installation Automatic Installation (Smithery) To install the Yonote MCP server for Claude Desktop: bash npx -y @smithery/cli install @cutalion/yonote-mcp --client claudeManual Installation 1. Clone the repository:bash git clone <your-repo-url> cd yonote-mcp 2. Install dependencies: Using uv: bash uv pip install -r requirements.txt # or uv pip install . 3. Configure Environment Variables: Create a .env file in the project root: env API_TOKEN=your_yonote_api_token API_BASE_URL=https://app.yonote.ru/api # Optional ---
2. Configuration Cursor Configuration Add the following to your ~/.cursor/mcp.json file (replace /path/to/yonote-mcp with your actual local path): json { "mcpServers": { "yonote": { "command": "uv", "args": [ "run", "-v", "--directory", "/path/to/yonote-mcp", "/path/to/yonote-mcp/main.py" ] } } } ---
3. Available Tools * documents_list: Get a list of documents. * Parameters: limit (optional), offset (optional), collectionId (optional). * documents_info: Get detailed information about a specific document. * Parameters: id (required). * collections_list: Get a list of collections. * Parameters: limit (optional), offset (optional). ---
4. Example Prompts * "List all my documents from Yonote." * "Show me the details for the Yonote document with ID [ID]." * "Fetch a list of my Yonote collections."
What you can do with Yonote
Use Case 1: Context-Aware Development in Cursor Problem: Developers often store internal documentation, architecture decisions, or coding standards in a Yonote wiki. When writing code in an IDE like Cursor, the AI doesn't have access to these private "source of truth" documents, leading to suggestions that might violate company standards or miss internal logic. Solution: By integrating the Yonote MCP with Cursor, the AI can proactively search your Yonote collections for relevant documentation. It can read the requirements or style guides directly from your workspace to ensure the code it generates aligns with your team's specific guidelines. Example: A developer asks the AI, "Implement the authentication logic based on our internal security document." The AI uses collections_list to find the "Engineering" folder, identifies the "Security-v2" document via documents_list, reads the content using documents_info, and then writes the code following those exact specifications.
Use Case 2: AI-Powered Knowledge Base Q&A Problem: As a Yonote workspace grows, finding specific information across dozens of collections and hundreds of documents becomes difficult. Traditional keyword search often fails to provide a synthesized answer from multiple sources. Solution: This MCP allows an AI assistant (like Claude) to act as a librarian for your Yonote data. The assistant can list documents across various collections, retrieve the full text of relevant entries, and provide a synthesized answer to complex questions. Example: A team member asks, "What are our current policies on remote work and hardware stipends?" The AI scans the "HR" and "Operations" collections, retrieves the relevant documents, and provides a concise summary of the policies instead of the user having to open and read four different pages.
Use Case 3: Automated Content Repurposing and Drafting Problem: Content creators and managers often store raw research, interview transcripts, or rough notes in Yonote. Moving this data manually to an AI tool to draft a blog post, newsletter, or report is a tedious "copy-paste" cycle. Solution: Using the Yonote MCP, a user can instruct an AI to draft a new piece of content based on existing notes directly. The AI can pull the latest research from a specific Yonote collection to ensure the draft is based on the most recent data. Example: A user tells Claude: "Look at the latest 3 documents in my 'Product Research' collection in Yonote and draft a 500-word summary for the stakeholders' weekly email." The AI fetches the documents using documents_list and…