FileForge — File Converter MCP Server

The open-source ilovepdf.com for AI agents and humans alike.
83 file conversion tools across 9 categories — usable as an MCP server by Claude, Cursor, and any MCP-compatible agent, or as a standalone web app.

What is this?

Most AI agents can write and read files, but they can't convert them. FileForge fills that gap.

Add it to Claude, Cursor, Windsurf, or any MCP client and your AI can:

"Compress this PDF and protect it with a password"
"Convert all images in this folder to WebP at 80% quality"
"Extract tables from this PDF to Excel"
"Turn this CSV into a Parquet file"
"Merge these three PDFs and add page numbers"
"OCR this scanned document and make it searchable"

No API keys. No cloud. Everything runs locally.

Features at a Glance

| Category | Tools | Highlights | |----------|-------|-----------| | PDF Core | 11 | Merge, split, compress, rotate, protect/unlock, watermark, page numbers, organize, repair | | PDF → Other | 7 | PDF → Word, Excel, PowerPoint, Images, HTML, Text, Markdown | | Other → PDF | 5 | Images, Word, Excel, PowerPoint, HTML → PDF | | OCR | 6 | Scanned PDF → searchable PDF, image → text/Word/HTML/Markdown, 80+ languages | | Images | 7 | PNG/JPG/WebP/BMP/GIF/TIFF conversion, resize, GIF frames, Base64, HEIC | | Documents | 20 | DOCX/HTML/Markdown/TXT/PPTX/XLSX → any other doc format | | Data | 18 | JSON/YAML/CSV/XML/TOML/XLSX/Parquet/SQLite/JSONL conversions | | Archives | 6 | Create & extract ZIP, TAR.GZ, 7Z | | E-books | 7 | EPUB → TXT/HTML/Markdown/DOCX/PDF, RTF → TXT/HTML |

Total: 87 tools (83 conversion tools + get_server_info)

Installation

git clone https://github.com/Nidhishpajni/file-converter-mcp.git
cd file-converter-mcp
pip install -r requirements.txt

Optional extras

# For HTML → PDF with full CSS support
pip install weasyprint

# For Office → PDF (also requires Microsoft Office on Windows)
pip install docx2pdf

# For HEIC/HEIF image support
pip install pillow-heif

# For e-book tools
pip install ebooklib striprtf

# For Parquet support
pip install pyarrow

# Pre-download OCR models (~200 MB)
python -c "import easyocr; easyocr.Reader(['en'])"

Quick Start

Add to Claude Code (CLI)

claude mcp add file-converter python /path/to/file-converter-mcp/file_converter_mcp_server.py

Add to Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "file-converter": {
      "command": "python",
      "args": ["/absolute/path/to/file-converter-mcp/file_converter_mcp_server.py"]
    }
  }
}

Add to Cursor / Windsurf / any MCP client

{
  "mcpServers": {
    "file-converter": {
      "command": "python",
      "args": ["/absolute/path/to/file-converter-mcp/file_converter_mcp_server.py"]
    }
  }
}

Run the Web UI

pip install fastapi uvicorn python-multipart
python web_app.py
# Open http://localhost:8080

Start the MCP server manually

python file_converter_mcp_server.py              # stdio (default)
python file_converter_mcp_server.py --transport sse --port 8000

Test with MCP Inspector

mcp dev file_converter_mcp_server.py

Full Tool Reference

PDF Core (11 tools)

| Tool | Description | |------|-------------| | merge_pdfs | Combine multiple PDFs into one | | split_pdf | Split by page ranges or extract every page | | compress_pdf | screen (72 DPI), ebook (150 DPI), prepress (lossless) | | rotate_pdf | Rotate all or selected pages 90°, 180°, or 270° | | protect_pdf | AES-256 password protection | | unlock_pdf | Remove password from a protected PDF | | add_watermark | Diagonal text or image watermark with opacity control | | add_page_numbers | Customisable position, font size, starting number | | organize_pdf | Reorder, duplicate, or delete pages | | repair_pdf | Fix corrupted or malformed PDFs | | get_pdf_info | Metadata, page count, dimensions, encryption status |

PDF → Other (7 tools)

| Tool | Description | |------|-------------| | pdf_to_images | Render pages as PNG/JPG/WebP at any DPI | | pdf_to_text | Extract plain text | | pdf_to_word | Convert to editable .docx | | pdf_to_excel | Extract tables to .xlsx | | pdf_to_pptx | Convert to .pptx (image or text mode) | | pdf_to_html | Convert to structured HTML | | pdf_to_markdown | Convert to Markdown |

Other → PDF (5 tools)

| Tool | Description | |------|-------------| | images_to_pdf | Combine images into a single PDF | | word_to_pdf | .docx → PDF (requires Microsoft Word on Windows) | | excel_to_pdf | .xlsx → PDF (requires Microsoft Excel on Windows) | | pptx_to_pdf | .pptx → PDF (requires Microsoft PowerPoint on Windows) | | html_to_pdf | HTML → PDF (WeasyPrint or ReportLab fallback) |

OCR (6 tools)

| Tool | Description | |------|-------------| | ocr_pdf | OCR scanned PDF; optionally overlay invisible text layer | | ocr_image | Extract text from any image (plain text or JSON with coords) | | image_to_word | OCR image → .docx | | image_to_html | OCR image → HTML | | image_to_markdown | OCR image → Markdown | | list_ocr_languages | List all 80+ supported EasyOCR language codes |

Images (7 tools)

| Tool | Description | |------|-------------| | convert_image | Convert between PNG, JPG, WebP, BMP, GIF, TIFF with optional resize | | get_image_info | Format, mode, dimensions, file size | | gif_to_frames | Extract every frame from an animated GIF | | frames_to_gif | Combine images into an animated GIF | | image_to_base64 | Encode an image as a Base64 data URI | | base64_to_image | Decode a Base64 string back to an image | | heic_to_image | Convert Apple HEIC/HEIF photos to JPG, PNG, or WebP |

Documents (20 tools)

| Input | Outputs | |-------|--------| | .docx | HTML, Markdown, TXT | | .html | DOCX, Markdown, TXT, XLSX | | .md | HTML, PDF, DOCX, TXT | | .txt | PDF, DOCX, HTML | | .pptx | TXT, HTML, Images | | .xlsx | HTML, Markdown, DOCX |

Data (18 tools)

| Tool | Description | |------|-------------| | convert_data | JSON ↔ YAML ↔ CSV ↔ TOML ↔ XLSX | | xml_to_json / json_to_xml | XML ↔ JSON | | xml_to_csv | Flatten XML to CSV | | ini_to_json / ini_to_yaml | INI config → JSON / YAML | | env_to_json | .env file → JSON | | csv_to_markdown | CSV → Markdown table | | html_table_to_csv | Extract HTML <table> → CSV | | jsonl_to_json / json_to_jsonl | JSONL ↔ JSON array | | csv_to_parquet / parquet_to_csv | CSV ↔ Parquet | | parquet_to_json | Parquet → JSON | | sqlite_to_csv / sqlite_to_json / sqlite_to_xlsx | SQLite → CSV/JSON/XLSX |

Archives (6 tools)

| Tool | Description | |------|-------------| | zip_files / unzip_files | Create and extract ZIP archives | | tar_files / untar_files | Create and extract TAR archives | | create_7z / extract_7z | Create and extract 7Z archives |

E-books (7 tools)

| Tool | Description | |------|-------------| | epub_to_txt/html/markdown/docx/pdf | Convert EPUB to various formats | | rtf_to_txt / rtf_to_html | Convert RTF documents |

Architecture

file-converter-mcp/
├── file_converter_mcp_server.py   # MCP server entry point (stdio / SSE)
├── web_app.py                     # FastAPI web server + drag-and-drop UI
├── static/index.html              # FileForge web UI (single file, no build step)
├── tools/                         # One module per category
└── utils/                         # Shared helpers (file paths, PDF, OCR)

All tools return {"success": True, ...} or {"success": False, "error": "..."} — never raise exceptions.

Limitations

| Area | Notes | |------|-------| | PDF → Word/Excel fidelity | Open-source can't match Adobe. Complex layouts need cleanup. | | Office → PDF | Requires Microsoft Office on Windows (COM automation). | | OCR models | First call downloads EasyOCR weights (~100–400 MB). | | HEIC | Requires pip install pillow-heif. | | E-books | Require pip install ebooklib striprtf. | | Audio/Video | Not supported (would need ffmpeg). |

Contributing

PRs welcome! High-value areas: Docker image, audio/video (ffmpeg), SVG support, better PDF→Word fidelity.

License

MIT — see LICENSE

If this saved you time, please ⭐ the repo — it helps others find it!

MCP Servers