Running Translation Servers

This guide teaches you how to set up and run the background servers required for the various translation engines in Cheat UI.

---

🏗️ Architecture Overview

The translation system works as follows:

1. Game (Cheat UI): Sends Japanese text to a local or remote server.
2. Background Server: Receives the text, translates it using its engine, and returns it.
3. Cheat UI: Updates the game data (Map names, Variables, etc.) with the translated text.

IMPORTANT

You must have the background server running BEFORE you click "Start Translation" in-game.

---

🇰🇷 ezTrans Setup (JP → KR)

For high-quality Japanese to Korean translation, you need the original ezTrans XP software and a "wrapper" server.

1. Prerequisite: ezTrans XP

You must have ezTrans XP installed on your Windows machine. It is a commercial product and must be set up correctly in your system registry.

2. Choose a Wrapper Server

Option A: ezTransWeb (Python)

• GitHub: HelloKS/ezTransWeb
• Setup:
  1. Install Python 3.
  2. Install Flask: pip install flask
  3. Download ezTransWeb.py from the repository.
  4. Run: python ezTransWeb.py
• Port: 5000 (Default).

Option B: eztrans-server (Rust/High Performance)

• GitHub: nanikit/eztrans-server
• Setup:
  1. Download the latest .exe from the Releases page.
  2. Run: Just double-click the eztrans-server.exe.
• Port: 8000 (Default).

---

🐳 Lingva Setup (Auto-detect → EN/Others)

Lingva is an open-source front-end for Google Translate. For bulk translations, you should always run it locally via Docker.

Running a Cluster (Recommended)

This repository includes a docker-compose.yml that runs a 3-node balanced cluster for extreme performance (~160 strings/sec).

1. Install Docker Desktop.
2. Open your terminal in the plugin root directory.
3. Run: docker-compose up -d
4. Cheat UI will now be able to use ports 3000, 3001, and 3002.

Running a Single Instance

If you just want a simple local node:

docker run -d -p 3000:3000 thedaviddelta/lingva-translate

---

🤖 AI/LLM Setup (Generic)

For the smartest translations, use an Artificial Intelligence model.

Ollama (Local AI)

1. Install Ollama.
2. Pull a model: ollama pull personal/llama3-8b-instruct (or any other translation model).
3. Run: The Ollama server runs automatically in the background.
4. In Cheat UI, select the Ollama (Local) preset.

---

🧪 Testing with Dummy Server

If you want to test the UI and connectivity without setting up a real engine, you can use the included dummy server:

1. Open terminal in dummy-translator/.
2. Run: python eztrans.py
3. In Cheat UI, select ezTransWeb as the endpoint.
4. It will return "T: [Original Text]" to confirm the connection is working.