👻 WRAITH: Resurrecting Lost Memories with Local Gemma 4 — The Ghost in the Machine

Description and use

WRAITH is a local-first AI memory archive that transforms historical letters, diary entries, and personal recollections into atmospheric gothic stories while keeping factual information separate from creative interpretation, powered by Gemma 4 through Ollama.

Problem

Personal diaries, handwritten letters, and family archives can hold irreplaceable memories, but preserving and exploring them with modern AI introduces privacy concerns, historical accuracy risks from generative AI inventing details, and accessibility barriers from cloud-dependent or paid-API solutions.

Outcome

WRAITH produces a dual-layer output: a Verified Ledger extracting explicitly mentioned people, dates, events, and other details from source text, and a Haunted Chronicle generating atmospheric gothic ghost stories inspired by the original text's mood, imagery, and emotional themes, with factual extraction kept separate from creative storytelling.

Workflow

• Download the Gemma 4 model via Ollama: ollama pull gemma4:e2b
• Clone the repository: git clone https://github.com/PruthviRajG25/Hacktoberfest-Week4.git
• Install dependencies: npm run install:all
• Start the application: npm run dev
• Open http://localhost:3000 in the browser

Evidence

WRAITH is a privacy-focused archival and storytelling application powered by Google DeepMind's Gemma 4, running locally through Ollama.

Source

The backend sends structured instructions to the model and requests a JSON response... Extract only names, dates, and events explicitly supported by the input... Keep factual extraction separate from creative storytelling.

Source

Every Memory Leaves a Shadow.

Source

You will need Node.js, npm, and Ollama installed.

Source