ReMind is an AI-powered personal memory agent designed to capture, transcribe, index, and summarize users' digital activities for easy recall. It targets individuals seeking to manage and query their digital history through an interactive chat interface, offering a local, on-device solution for enhanced privacy.

How It Works

ReMind captures digital activities including screenshots and audio, transcribing text from images. It leverages vector databases for indexing this data, enabling efficient retrieval. Large language models (LLMs) are used for summarizing daily activities and powering the conversational query interface, allowing users to ask questions about their captured digital history.

Quick Start & Requirements

• Install: git clone https://github.com/DonTizi/ReMind/, then npm install and npm run dev.
• Prerequisites: Node.js (18+), npm, Ollama (with llama3.1 pulled), Python.
• Setup: Initial indexing may take ~30 minutes. A macOS version is available.
• Resources: Full setup video available.

Highlighted Details

• Captures screenshots and audio.
• Transcribes text from screenshots.
• Indexes data using vector databases.
• Provides daily activity summaries.
• Features an interactive chat interface for querying history.

Maintenance & Community

• Community contributions are welcomed via pull requests.
• Contact via GitHub issues or elyes.melbouci@gmail.com.
• Roadmap available for viewing.

Licensing & Compatibility

• Licensed under the Apache License.
• Compatible with commercial use and closed-source linking.

Limitations & Caveats

The project is currently distributed as a macOS application, with other platforms not explicitly mentioned. Initial setup requires downloading and configuring multiple external dependencies including Ollama and Python.