mcp

This MCP server is designed to generate outlines and presentations for Cappt using the Model Context Protocol. It exposes tools such as generate_outline to create a standard outline from user input, and generate_presentation to build a full presentation from an outline with optional gallery and preview content. To use it, first generate an outline by passing your input (title, article, etc.) to the generate_outline tool. Once you have a suitable outline, call generate_presentation with the outline and set preferences for include_gallery and include_preview to tailor the output. The response will include a record_id for tracking, a status, total_page count, an editor_url for quick edits, and the resulting presentation metadata like title, thumbnail, gallery, and preview. You can adjust the workflow by toggling the include_gallery and include_preview flags to suit your presentation needs.

Prerequisites:

• Node.js and npm installed on your machine
• Access to the MCP repository this server belongs to
• Optional: CAPPT_TOKEN if you plan to enable authenticated features

Installation steps:

1. Clone or download the repository git clone https://github.com/cappt-team/mcp-server-cappt cd mcp-server-cappt

2. Install dependencies npm install

3. Build the server (if a build step exists in your project) npm run build

4. Run the MCP server locally (example) npm start

5. Configure the server in your MCP environment using the provided mcp_config snippet (see the documentation above) and replace PLACEHOLDER_TOKEN with your actual token if required.

• The CAPPT_TOKEN environment variable is required for authenticated operations; replace YOUR_CAPPT_TOKEN with your real token in production.
• The server exposes two primary tools: generate_outline (to produce a standard outline from input) and generate_presentation (to build a presentation from an outline).
• When using generate_presentation, you can control gallery and preview inclusion via include_gallery and include_preview flags.
• Ensure dist/index.js (or the built entry point) is correctly referenced in your deployment configuration.
• If you encounter issues with the MCP inspector, ensure you have installed the inspector package via uv as described in the repository's debugging guidance.