Document Converter API

Lightning-fast, cloud-native document conversion powered by GraalVM Native Image.

Transform documents with sub-second startup and 50% less memory footprint. Choose from three specialized engines for your specific needs.

🚀 Quick Start with Docker

# MarkItDown - Office docs to Markdown
 docker pull molandev/markitdown-converter-api:latest

# LibreOffice - Universal format conversion
 docker pull molandev/libreoffice-converter-api:latest

# MinerU - PDF parsing with OCR
 docker pull molandev/mineru-converter-api:latest

Run in Seconds

docker run -d -p 10996:10996 molandev/markitdown-converter-api:latest
# That's it! Visit http://localhost:10996/help for API docs

📦 Choose Your Engine

Engine	Best For	Pull Command
MarkItDown	DOCX/PPTX/XLSX → Markdown	docker pull molandev/markitdown-converter-api:latest
LibreOffice	100+ formats (DOC→DOCX, DOCX→PDF, etc.)	docker pull molandev/libreoffice-converter-api:latest
MinerU	PDF → Markdown/JSON (with OCR)	docker pull molandev/mineru-converter-api:latest

📋 Detailed Overview

• MarkItDown Overview - Office documents to Markdown
• LibreOffice Overview - Universal format conversion
• MinerU Overview - PDF parsing with OCR

✨ Why Choose This?

✅ Zero JVM Warmup - GraalVM Native Image means instant peak performance
✅ Resource-Efficient - Runs comfortably on 2C4G instances
✅ Production-Ready - Smart concurrency control & auto-cleanup
✅ Cloud-Native - Docker-first design, Kubernetes-ready
✅ Chinese Filename Support - URL-encoded headers for international compatibility

🔥 Core Features

Smart Concurrency Control

Semaphore-based rate limiting prevents system overload:

• Configurable max concurrent conversions
• Graceful queuing with timeout
• Automatic "system busy" response when overloaded

Auto-Cleanup Mechanism

Scheduled task automatically purges temporary files:

• Configurable retention time
• Configurable cleanup interval
• Prevents disk space exhaustion

Memory Optimized

Stream-based file handling:

• No OOM errors with large files (100MB+)
• Low memory footprint
• Efficient resource utilization

📚 API Examples

MarkItDown (Office → Markdown)

# Convert Word document
 curl -F "file=@report.docx" http://localhost:10996/convert/markitdown/upload -o report.md

# Keep embedded images
 curl -F "file=@presentation.pptx" -F "keepDataUris=true" http://localhost:10996/convert/markitdown/upload -o output.md

LibreOffice (Universal Conversion)

# DOCX to PDF
 curl -F "file=@document.docx" -F "format=pdf" http://localhost:10996/convert/libreoffice/upload -o output.pdf

# DOC to DOCX
 curl -F "file=@legacy.doc" -F "format=docx" http://localhost:10996/convert/libreoffice/upload -o output.docx

MinerU (PDF → Markdown/JSON)

# Get Markdown
 curl -F "file=@document.pdf" http://localhost:10996/convert/mineru/upload -o output.md

# Get all files (ZIP)
 curl -F "file=@document.pdf" "http://localhost:10996/convert/mineru/upload?outputFormat=zip" -o output.zip

⚙️ Configuration

Environment Variable	Default	Description
CONVERTER_MAX_CONCURRENT	10	Max parallel conversions
CONVERTER_RETENTION_MINUTES	10	Temp file retention (minutes)
CONVERTER_SCHEDULE_MINUTES	5	Cleanup interval (minutes)
CONVERTER_TEMP_DIR	/app/temp	Temp file directory
SERVER_PORT	10996	Service port

Performance Tuning

# High-concurrency setup
docker run -e CONVERTER_MAX_CONCURRENT=20 \
           -e CONVERTER_ACQUIRE_TIMEOUT=60 \
           molandev/markitdown-converter-api:latest

# Resource-constrained environment
docker run -e CONVERTER_MAX_CONCURRENT=3 \
           -e CONVERTER_RETENTION_MINUTES=5 \
           molandev/markitdown-converter-api:latest

# With persistent temp directory
docker run -v /data/converter-temp:/app/temp \
           -e CONVERTER_RETENTION_MINUTES=30 \
           molandev/markitdown-converter-api:latest

🎯 Perfect For

• AI Knowledge Base Pipelines - Convert documents to LLM-friendly Markdown
• Document Management Systems - Batch conversion workflows
• Content Migration Projects - Legacy format modernization
• SaaS Applications - Multi-tenant document processing
• Research & Analysis - Extract structured content from PDFs

🏗️ Build from Source

# MarkItDown
 docker build -f Dockerfile-markitdown -t molandev/markitdown-converter-api:latest .

# LibreOffice
 docker build -f Dockerfile-libreoffice -t molandev/libreoffice-converter-api:latest .

# MinerU
 docker build -f Dockerfile-mineru-pipline -t molandev/mineru-converter-api:latest .

📖 Documentation

• API Documentation: http://localhost:10996/help (available after starting the service)
• 中文文档: README-zh.md

---

Built with Spring Boot 3.5.6 + GraalVM 25

One service, three engines, infinite possibilities. ⚡