Ghost Papers: an offline AI that reads and explains the papers a family leaves behind

Description and use

Ghost Papers helps families organize and understand personal documents left behind after a loved one’s death or illness, entirely offline and in their native language.

Problem

Families struggle to understand and organize the diverse, private documents left behind after a loved one’s death or illness.

Outcome

Ghost Papers enables families to inventory and understand private documents offline, preserving privacy and providing actionable next steps in their native language.

Workflow

• Drop photos or PDFs of papers
• Gemma 3 (4B) via Ollama performs OCR locally
• Second Gemma call converts extracted text to strict JSON
• Code validates that every name, number, and date appears in the OCR text
• Numbers and dates are normalised and masked
• Receipts link each fact to the exact OCR line
• Legacy file and printable binder are generated
• Optional local storage with masked identifiers is saved
• Scrub step removes any raw identifiers before writing

Evidence

Ghost Papers reads those papers, explains each one in plain English, বাংলা or हिन्दी, and files everything into one legacy file

Source

Gemma 3 (4B) through Ollama, running locally. Gemma's vision reads each picture, and a second call turns the text into strict JSON. No API key, no cloud, $0.

Source

Every name, number and date must be found in the text that was read, or it is dropped, and the app shows how many claims it blocked.

Source

Privacy: cases live in memory only. Pictures are read, then dropped. Saving is opt-in, local, and stores masked identifiers with no pictures and no full text.

Source