MCP Server for AI Agents

FilesToText MCP (Model Context Protocol) server for AI agents and LLMs to convert files to Markdown.

🤖 FilesToText MCP Server

The FilesToText MCP server enables AI agents and large language models to convert files to Markdown directly within their workflows.

Quick Start

MCP Server Endpoint: filestotext.com/mcp

Protocol: Model Context Protocol (MCP)

Authentication: None required (public server)

What is MCP?

The Model Context Protocol (MCP) is an open standard that allows AI models and agents to interact with external tools and data sources. FilesToText provides an MCP server that integrates file conversion capabilities into your AI workflows.

Available Tools

1. Convert File to Markdown

Convert any supported file format to clean, structured Markdown.

Tool Name: convert_file

Parameters:

  • file_name (string, required) - file name, used to detect conversion type
  • file_content (string, optional) - Base64-encoded file content

Returns:

{
  "markdown": "# Converted content...",
  "success": true,
  "format": "pdf",
  "file_size": 1024
}

2. Get Supported Formats

Retrieve the list of all supported file formats.

Tool Name: get_supported_formats

Parameters: None

Returns:

{
  "formats": [
    {
      "extension": ".pdf",
      "mimeType": "application/pdf",
      "name": "PDF Document"
    },
    ...
  ],
  "total": 42
}

Use Cases

📄 Document Analysis

Let AI agents analyze documents by converting them to Markdown first:

User: "Analyze this contract and highlight key terms"
→ Agent converts PDF to Markdown
→ Agent analyzes the Markdown content
→ Agent returns summary

📊 Data Extraction

Extract structured data from various file formats:

User: "Extract all tables from this spreadsheet"
→ Agent converts XLSX to Markdown
→ Agent parses Markdown tables
→ Agent returns structured data

🔍 Content Processing

Process and transform content across formats:

User: "Convert this presentation to a blog post"
→ Agent converts PPTX to Markdown
→ Agent rewrites as blog post
→ Agent returns formatted content

🤝 Multi-Format Workflows

Handle mixed file types in a single workflow:

User: "Summarize all these documents"
→ Agent converts each file (PDF, DOCX, XLSX)
→ Agent summarizes each
→ Agent creates unified summary

Features

✅ No Authentication - Public server, no API keys needed ✅ Privacy Guaranteed - Files processed in-memory, never stored ✅ Fast Processing - Conversion typically < 2 seconds ✅ 10+ Formats - PDF, DOCX, XLSX, CSV, and more ✅ Error Handling - Detailed error messages for debugging ✅ Rate Limiting - 10 requests/minute

Rate Limits

  • Free Tier: 10 requests/minute
  • File Size: Max 4 MB per request
  • Timeout: 30 seconds per request

Error Handling

No file provided:

{
  "error": "No file provided",
  "success": false
}

File too large:

{
  "error": "File size exceeds 4 MB limit",
  "success": false
}

Conversion failed:

{
  "error": "Conversion failed",
  "success": false
}

Privacy & Security

  • No Storage: Files processed in-memory only
  • No Logging: File contents never logged
  • Immediate Deletion: All data deleted after conversion
  • HTTPS Only: All traffic encrypted
  • No Tracking: No user data collected

Configuration

Add to your mcp.json:

{
  "mcpServers": {
    "files-to-text": {
      "url": "https://filestotext.com/mcp"
    }
  }
}

Advertisement