PDF-2-Markdown | System Prompt Improvements

PDF-2-Markdown v1.2.0 (Windows Release)

This release packs the entire tool into a single, standalone .exe file for Windows users. You don't need Python or any external packages installed to run it—just download it and you're good to go!

Details

• No Python Required: Runs straight out of the box as a single .exe file.
• Flexible API Support: Works with Google Gemini models as well as local setups (like LM Studio).
• Batch Conversion: Process a single PDF or convert every PDF in your folder all at once.
• Custom DPI Settings: Adjust resolution (100 to 1000 DPI) to get the clearest OCR results for your documents.
• Clean Formatting: Extracts text while preserving tables, math equations (in LaTeX), and clickable links.
• System Prompt: Better overall output accuracy

---

Test Run & Performance

Original PDF

Markdown Output

Quick Start Guide - Try yourself

1. Download pdf2markdown.exe from the Assets section below.
2. Drop the .exe file into the folder that contains your PDF files.
3. (Optional) Create a .env file in the same folder to automatically load your API key:

Google_API_KEY = your_actual_api_key_here

5. Double-click pdf2markdown.exe (or run it from your terminal) and follow the simple on-screen instructions.

---

Feel free to open an issue if you run into any bugs or have suggestions.

PDF-2-Markdown v1.1.0 - Standalone Executable

PDF-2-Markdown v1.1.0 (Windows Release)

Hey there! This release packs the entire tool into a single, standalone .exe file for Windows users. You don't need Python or any external packages installed to run it—just download it and you're good to go!

Details

• No Python Required: Runs straight out of the box as a single .exe file.
• Flexible API Support: Works with Google Gemini models as well as local setups (like LM Studio).
• Batch Conversion: Process a single PDF or convert every PDF in your folder all at once.
• Custom DPI Settings: Adjust resolution (1 to 600 DPI) to get the clearest OCR results for your documents.
• Clean Formatting: Extracts text while preserving tables, math equations (in LaTeX), and clickable links.

---

Test Run & Performance

Original PDF

Markdown Output

Quick Start Guide - Try yourself

1. Download pdf2markdown.exe from the Assets section below.
2. Drop the .exe file into the folder that contains your PDF files.
3. (Optional) Create a .env file in the same folder to automatically load your API key:

Google_API_KEY = your_actual_api_key_here

5. Double-click pdf2markdown.exe (or run it from your terminal) and follow the simple on-screen instructions.

---

Feel free to open an issue if you run into any bugs or have suggestions.

Full Changelog: v1.0.0...v1.1.0

v1.0.0 - Initial Release: RAG-Ready PDF to Markdown Converter

Welcome to the first official release of PDF 2 Markdown!

This CLI utility leverages Vision Large Language Models to convert complex PDF documents into clean, structured .md files, preserving the original layout, paragraphs, and math formulas. It's designed to be the perfect first step for feeding documents into local RAG (Retrieval-Augmented Generation) pipelines.

Key Features in v1.0.0

• Multi-Model Support: Seamlessly integrates with Google's Gemini API (Flash/Pro) and supports offline, local models via LM Studio.
• Flawless Formatting: Strictly extracts raw text without AI hallucinations or Markdown code blocks.
• Advanced Math Support: Automatically detects and formats mathematical equations using strict $ and $$ LaTeX syntax.
• Token Optimized: Converts PDFs to Grayscale PNGs and Base64 internally to save on API payload size.
• Interactive CLI: A simple, guided terminal interface to select your files, APIs, and DPI.

How to Use (Windows Standalone)

No Python environment required!

1. Download the pdf2markdown.exe file attached below.
2. Place it in the same folder as your .pdf files.
3. Run the executable and follow the terminal prompts.

(Note: If using the Gemini API, you will be prompted to paste your API key on the first run).

For source code usage and developer setup, please refer to the README.md.