Memora：找回遺失的截圖

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

Memora is an AI-powered screenshot search tool that helps students, researchers, and developers find information buried in their screenshots using natural-language queries. Instead of manually scrolling through hundreds of images, users can search for concepts, diagrams, code snippets, charts, and notes.

Memora combines OCR, AI-powered screenshot understanding, and semantic search to make screenshots searchable by meaning, not just exact text. It uses open-weight AI models locally, helping keep personal data private and reducing dependence on paid cloud APIs.

Students, researchers, developers, and professionals who frequently save screenshots could pay for Memora to save time, stay organized, and retrieve important information instantly.

Demo

Code

MemoryBox web demo

MemoryBox is a privacy-first screenshot search demo. Drop screenshots into the browser, let local Tesseract.js OCR read their text, then search that extracted text. Optionally, use a locally running Ollama vision model to create extra screenshot metadata. There is no application backend, account system, analytics, upload endpoint, cloud OCR, or cloud AI API.

Run locally

Use Node.js 20 or newer:

npm install
npm run dev

Open the local Vite address shown in the terminal (normally http://localhost:5173). For a production build:

npm run typecheck
npm test
npm run build
npm run preview

Demo workflow

1. Drag PNG, JPG/JPEG, or WebP screenshots onto the upload panel, or use Choose images.
2. Approve the browser’s first-time OCR language-data download if prompted by Tesseract.js.
3. Watch per-image indexing progress and failures.
4. Optionally open Local AI and use Test connection. MemoryBox defaults to gemma3:4b at http://localhost:11434.
5. Enable automatic analysis for new…

How I Built It

I built Memora as a browser-based application using JavaScript, with OCR-powered text extraction, IndexedDB for local persistence, and a hybrid search system combining keyword-based and semantic retrieval.

I integrated Gemma 3:4B through Ollama for local AI-powered screenshot understanding and metadata generation, along with nomic-embed-text for generating embeddings and finding screenshots based on semantic similarity. Tesseract.js extracts text from uploaded screenshots, while AI-generated descriptions, categories, keywords, and embeddings help users discover images even when their queries don't exactly match the text.

The project is designed around open-weight models and local inference, keeping screenshot data on the user's device and reducing reliance on external AI APIs. This makes Memora a privacy-focused, AI-powered visual memory tool.

Why Does Open Innovation Matter?

Open innovation made it possible to build Memora around locally running AI models instead of relying entirely on closed, paid APIs. Using Gemma 3:4B through Ollama gave me control over how screenshots are analyzed, while nomic-embed-text enabled semantic search.

Unlike a closed API-dependent approach, this architecture offers greater control over data privacy, model usage, and running costs. It also lets me experiment, customize the pipeline, and improve the product without being locked into a single provider. Open-weight models made building a privacy-focused AI tool more accessible as a student developer with limited resources.

My First User

My first users were my college peers, who tried Memora for free and found it useful enough to purchase a one-time subscription for ₹100. Seeing people move from trying my product to paying for it was an encouraging milestone. It gave me confidence that Memora could solve a real problem and motivated me to keep improving it.

Prize Categories

Best Use of Gemma — Used Google's open-weight Gemma 3:4B model through Ollama for local AI-powered screenshot understanding and metadata generation in Memora.

Thank You!