Quick Start

TransLite is a lightweight desktop translation app. Follow these steps to get up and running in minutes.

Prerequisites

  • Node.js 18+ and npm
  • Operating system: Windows 10+, macOS 12+, or Linux
  • For local translation: ~4GB free disk space for model download

Installation

Download Prebuilt Binary

Download the latest release for your platform from the Download page.

Build from Source

$ git clone https://github.com/yourusername/translite.git

$ cd translite

$ npm install

$ npm run build

$ npm start

Development Mode

$ npm run dev

This starts the Vite dev server for frontend hot-reload. Run npm start in another terminal to launch the Electron app pointing at the dev server.

Configuration

TransLite stores all settings locally in a JSON file at userData/settings.json. No backend or database required — your data stays on your machine.

Settings Overview

  • Provider configurations (Base URL, API Key, model selection)
  • Local model directory and model loading settings
  • Global shortcut key binding
  • Usage check templates (JavaScript request/extractor)
  • Theme and display preferences

Provider Setup

TransLite comes with presets for major LLM providers. Each provider can be independently configured.

Adding a Provider

  1. Open TransLite and navigate to Settings
  2. Select a provider from the preset list
  3. Enter your API Key and Base URL (if needed)
  4. Click “Fetch Models” to load available models
  5. Select a model and activate the provider

Provider Features

  • Endpoint Testing — Verify your API endpoint before using it
  • Model Fetching — Automatically retrieve available models from the provider
  • Per-Provider Activation — Enable or disable providers independently

OpenAI-Compatible Gateways

TransLite supports any OpenAI-compatible API gateway. Simply enter your custom Base URL and API Key, and the app will work with any provider that implements the OpenAI API format.

Local Models

TransLite uses ONNX Runtime for local model inference, enabling fully offline translation.

Default Model

The default local model is onnx-community/HY-MT1.5-1.8B-ONNX, which provides fast and accurate translation without an internet connection.

Privacy: Local model inference means your text never leaves your computer. All processing happens on-device.

Custom Models

You can use custom model directories. Download compatible ONNX models and point TransLite to your model folder in the settings. The app handles model loading and management automatically.

Keyboard Shortcuts

Quick-open translation windowAlt + Space
Close windowEsc
(More shortcuts coming soon)In development

The global shortcut is customizable in Settings.

Usage Tracking

TransLite provides a flexible usage tracking system using JavaScript templates. You can create custom request and extractor scripts to monitor API consumption for any provider.

How It Works

  • Request Template — Define how to query the provider's usage API
  • Extractor Template — Parse the API response to extract balance, token usage, and plan status
  • Usage is displayed in-app: balances, Token Plan, API availability, and per-request token counts

Frequently Asked Questions

Is TransLite free?

Yes! TransLite is open source under the MIT License. It's completely free to use and modify.

Do I need an internet connection?

No. TransLite supports fully offline translation using local ONNX models. Online LLM providers are optional.

How do I get an API key for providers?

Each provider has its own registration process. Visit the provider's website to sign up and get your API key.

Can I use my own local model?

Yes, you can configure a custom model directory in settings. The model must be in ONNX format.

Where are my settings stored?

Settings are stored locally at Electron's userData/settings.json. No data is sent to any server.

Which platforms are supported?

Windows 10+, macOS 12+, and Linux (modern distributions).