Update README.md

README.md

@@ -1,25 +1,28 @@
 ---
-title: Olmocr Demo
-emoji: 😻
-colorFrom: red
-colorTo: red
-sdk: docker
-sdk_version: 5.29.0
+title: olmOCR Markdown Converter
+emoji: 📝
+colorFrom: yellow
+colorTo: blue
+sdk: gradio
+sdk_version: 3.50.2
 app_file: app.py
-pinned: false
+python_version: 3.11
+license: mit
 ---
 
-# PDF to EPUB Converter (olmOCR)
+# olmOCR Markdown Converter
 
-This Gradio app converts a PDF into a clean EPUB using the [olmOCR](https://huggingface.co/allenai/olmOCR-7B-0225-preview) model. Each PDF page is rendered and processed through OCR, with the first page used as the cover. Metadata (title, author, language) can be entered manually.
+This Space uses the `olmOCR` model pipeline to convert PDFs (including scientific papers) into markdown `.txt` files that retain document structure, headers, and basic math formatting — ready for Calibre/Kindle or downstream parsing.
 
-## Features
+- ✅ Vision + text anchor OCR pipeline (via `olmOCR`)
+- ✅ Extracts semantic structure via PDF TOC
+- ✅ Outputs clean `.txt` in markdown format
+- ✅ Hugging Face **Gradio Space with GPU support**
 
-- OCR via `olmOCR-7B-0225-preview`
-- First page used as EPUB cover
-- Input for title, author, and language
-- EPUB output for ebook readers
+## Example Use
 
-## Requirements
+Upload a scientific paper in PDF and download a markdown `.txt` version with preserved headers and inline structure.
 
-Already defined in `requirements.txt`:
+---
+
+Built by [@BenedictRichardLeonardi](https://huggingface.co/BenedictRichardLeonardi) using [olmOCR](https://huggingface.co/allenai/olmOCR-7B-0225-preview)