Prism is a cross-platform AI translation application designed to provide accurate, context-aware translations and extract text from screenshots using advanced AI models and OCR technology. It targets users who require efficient and reliable translation tools on their desktop, offering a user-friendly interface and robust backend for seamless operation. The primary benefit is a powerful, integrated translation and OCR solution accessible across major operating systems.

How It Works

Prism employs a modern architecture combining a Vue 3 frontend with a Rust backend via the Tauri framework, ensuring native performance and security. Core translation capabilities are powered by Tencent's Hunyuan-MT-7B model, while integrated OCR recognition utilizes the Qwen3-VL-8B-Instruct model to extract text from images. Translations are processed via an OpenAI-compatible API provider, with users encouraged to use SiliconFlow for free credits. Translation history is stored locally using SQLite, ensuring user data privacy.

Quick Start & Requirements

To run Prism, clone the repository and install dependencies using pnpm install. Development mode can be started with pnpm tauri dev, and production builds with pnpm tauri build. Prerequisites include Rust (1.91.0+), Node.js (18+), pnpm, and Git. A crucial requirement is an OpenAI-compatible API key and base URL; users can obtain free credits via a SiliconFlow invite link.

Highlighted Details

• Cross-Platform Support: Fully functional on Windows and macOS, with Linux support listed as "Coming Soon."
• Advanced AI Models: Leverages Tencent Hunyuan-MT-7B for enterprise-grade translation and Qwen3-VL-8B-Instruct for sophisticated OCR.
• Integrated OCR: Enables direct text extraction and translation from screenshots.
• Local Data Storage: Translation history is saved locally via SQLite, safeguarding user privacy.

Maintenance & Community

The project is developed by the "Prism Team" and welcomes contributions via GitHub Issues and pull requests. While specific community channels like Discord or Slack are not listed, the roadmap indicates ongoing development with features like translation memory, batch file translation, and a plugin ecosystem planned.

Licensing & Compatibility

Prism is licensed under the MIT License, which is permissive and generally compatible with commercial use and closed-source applications.

Limitations & Caveats

Linux support is currently pending. The global hotkey feature is still under development. Online models require an internet connection; offline usage is only possible with locally hosted models. macOS users may encounter Gatekeeper warnings requiring specific terminal commands to bypass.