An offline assistant that reads your documents with Gemma 4 on your own GPU, then builds upload-ready files to each portal's exact rules.
What I Built If you've applied for a scholarship, an admission or an exam in India, you know this screen: Photograph: JPG, 3.5 cm × 4.5 cm, white background, 20–50 KB. Signature: JPG, 10–20 KB. Class 12 marksheet: PDF, front and back in one file, max 200 KB. Income certificate: issued on or after 01-04-2026, PDF, not more than 0.3 MB. Every portal wants the same documents in a slightly different…
Kaagaz is an offline tool that helps users prepare their documents for various scholarship, admission, and exam portals in India. Instead of uploading documents to external websites, Kaagaz processes them locally on the user's computer using Gemma 4 E4B running through Ollama. The tool reads the actual documents rather than relying on filenames, identifies each document type, and creates an editable checklist for the portal's specific requirements.
It ensures the documents are in the correct format, size, and date range before converting them to the required file type. Kaagaz also measures every output file to make sure they meet the portal's exact specifications. If a file fails to meet the criteria, it will not be saved, preventing accidental uploads. The entire process is built using Python, Pydantic, Pillow, PyMuPDF, OpenCV, SQLite, and Streamlit.
Since the user's GPU didn't meet the requirements for the larger 12B model, Kaagaz is optimized for the 4.3B E4B model, with any uncertain documents sent to a review queue for human verification. This offline solution saves time, ensures privacy, and helps users avoid issues caused by missing or incorrect information.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
