Report Generator MCP

A powerful CLI tool and MCP server for generating HTML reports with embedded charts and images.

Features

• Generate professional HTML reports from Markdown documents
• Embed interactive charts (bar, line, pie, doughnut, radar, polarArea)
• Embed AI-generated images using Pollinations.ai
• Embed images from URLs
• Edit existing reports by adding, removing, or replacing content
• MCP server integration for use with Claude Desktop

Installation

npm install -g @vint.tri/report_gen_mcp@latest

For Python dependencies (required for image generation/editing):

npm run install-python-deps

Or set up a Python virtual environment:

npm run setup-python

Environment Variables

• REPORTS_DIR: Directory where reports will be generated (required for CLI and HTTP API modes)
• CHUTES_API_TOKEN: API token for Chutes AI services (required for image generation/editing)
• IMG_SAVE_BASE_DIR: Base directory for saving generated/edited images

Usage

CLI Mode

# Set the reports directory
export REPORTS_DIR=/path/to/reports

# Generate a report
report_gen_mcp generate --document report.md --elements '{"chart1":{"type":"bar","config":{"labels":["A","B","C"],"datasets":[{"label":"Dataset 1","data":[1,2,3]}]}}}'

Programmatic Usage (Node.js)

import { generateReport, generateImage } from '@vint.tri/report_gen_mcp';

// Generate a report
const reportResult = await generateReport(
  '# My Report\nThis is a sample report with a chart.\n[[chart:chart1]]',
  {
    chart1: {
      type: 'bar',
      config: {
        labels: ['A', 'B', 'C'],
        datasets: [{
          label: 'Dataset 1',
          data: [1, 2, 3]
        }]
      }
    }
  },
  '/path/to/output/report.html'
);

// Generate an image
const imageDataUri = await generateImage('A beautiful landscape', {
  width: 512,
  height: 512,
  guidance_scale: 7.5
});

HTTP API Mode

# Start the server
report_gen_mcp start-server

# Generate a report via HTTP POST to http://localhost:3000/generate-report
curl -X POST http://localhost:3000/generate-report \
  -H "Content-Type: application/json" \
  -d '{
    "document": "# My Report\nThis is a sample report with a chart.\n[[chart:chart1]]",
    "elements": {
      "chart1": {
        "type": "bar",
        "config": {
          "labels": ["A", "B", "C"],
          "datasets": [{
            "label": "Dataset 1",
            "data": [1, 2, 3]
          }]
        }
      }
    }
  }'

MCP Server Mode (for Claude Desktop)

Run without arguments to start the MCP server:

report_gen_mcp

Available tools:

1. generate-report: Generate an HTML report with embedded charts and images
2. edit-report: Edit an existing HTML report by adding or removing content
3. generate-image: Generate an image from a text prompt using AI

Image Generation and Editing

This tool includes advanced image generation and editing capabilities powered by Chutes AI:

Generate Image

Create new images from text prompts with customizable parameters:

• Prompt (in English)
• Width and height (128-2048 pixels)
• Model selection
• Guidance scale
• Negative prompts
• Number of inference steps
• Seed for reproducibility

Edit Image

Modify existing images based on text descriptions:

• Input image path
• Output image path
• Edit prompt (in English)
• Width and height (512-2048 pixels)
• CFG scale
• Negative prompts
• Number of inference steps
• Seed for reproducibility

Report Elements

Reports can include various elements referenced in the Markdown document using placeholders like [[chart:id]] or [[image:id]].

Charts

• Bar charts
• Line charts
• Pie charts
• Doughnut charts
• Radar charts
• Polar area charts

Images

• AI-generated images (Pollinations.ai)
• Images from URLs

Examples

See the test files in the repository for examples of how to use all features.

License

MIT