mcp

This MCP server implements the Model Context Protocol (MCP) for interacting with BaseHub content and tooling. It exposes a suite of tools discovered from the src/tools directory, allowing an AI agent to create, update, and retrieve content blocks, manage assets, and automate workflows through natural language prompts. The server runs with two transport methods (SSE and STDIO) to accommodate different integration setups, so you can stream tool results or receive structured responses as the AI processes prompts. You can reference the documentation linked in the repository to understand how to prompt the agent to perform tasks such as adding content blocks, migrating existing data, or triggering forms and automations via the available tools.

To use it within your AI tooling stack, point your MCP-enabled agent at the running server and allow it to enumerate the available tools defined under src/tools. Each tool has a schema that describes its required parameters and a metadata object with a human-friendly name and description. The agent will select and invoke these tools based on your prompts, returning content structured for rendering in your UI or for further processing in your automation pipelines.

Prerequisites:

• Node.js (LTS version) and npm or pnpm/yarn
• Git
• Access to the project repository (clone or download)

1. Install dependencies

• Ensure you have Node.js and npm installed. You can verify with:
  • node -v
  • npm -v

2. Clone the repository

• git clone https://github.com/your-org/basehub-mcp.git
• cd basehub-mcp

3. Install project dependencies

• npm install

4. Run the development server

• npm run dev This starts the MCP server and enables both SSE and STDIO transport methods for integration.

5. Optional: build for production

• npm run build
• npm run start-sse or npm run start-stdio

Notes:

• If you use a different package manager, adapt the install and run commands accordingly (e.g., yarn install and yarn dev, or pnpm install and pnpm dev).
• The project discovers tools from src/tools automatically, so no extra configuration is typically required to enable the tooling surface.

Tips and considerations:

• The MCP server exposes tools defined under src/tools. Each tool includes a Zod-based schema and metadata to describe parameters and behavior.
• Transport methods: SSE is suitable for streaming long-running tool results, while STDIO can be easier to integrate in traditional CLI workflows.
• When deploying in production, consider locking tool schemas and enabling audit logs for tool invocations to aid debugging and governance.
• Ensure environment-specific variables (e.g., authentication tokens, API keys) are configured via env vars as needed; there are no mandatory env vars shown in the README, but your deployment may require them depending on the tools you enable.
• If you add new tools, remember to export a schema, metadata, and a default function implementation following the existing pattern in src/tools.