Voice-Powered Language Translator

A beginner-friendly desktop app for translating spoken or typed text between dozens of languages, powered by local AI (Ollama Llama 3.2) and a simple, beautiful interface.

---

Features

• Speak or type to translate between 30+ languages
• Simple, modern, and accessible UI
• Visual timer while recording
• Clear status feedback (Ready, Recording, Processing, Error, Done)
• All actions (Record, Stop, Clear, Translate) are one-click
• No internet required for translation (uses local Ollama model)
• Logs all sessions for review

---

Prerequisites

1. Python 3.8+

• Download Python

2. Required Python Packages

• speech_recognition
• requests
• tkinter (comes with most Python installations)
• pyaudio (for microphone input)

Install them with:

pip install speechrecognition requests pyaudio

On macOS, you may need PortAudio:

brew install portaudio
pip install pyaudio

3. Ollama (for local AI translation)

• Install Ollama
• Pull the Llama 3.2 model:

ollama pull llama3.2

• Start the Ollama server:

ollama serve

---

How to Run the App

1. Make sure Ollama is running (ollama serve in a terminal).
2. In another terminal, run:

python langAssist.py

---

Usage Instructions

1. Select source and target languages from the dropdowns.
2. To speak:
   • Click Record and speak your phrase.
   • Click Stop when finished.
   • The app will transcribe and translate your speech.
3. To type:
   • Type your text in the Transcript box.
   • Click Translate.
4. Clear resets everything for a new translation.
5. The Translation box will show your result.
6. The status bar at the bottom shows what the app is doing.

---

Troubleshooting

• Microphone not working?
  • Make sure your mic is plugged in and not muted.
  • On macOS, you may need to grant Terminal or Python access to the microphone in System Settings > Privacy & Security > Microphone.
  • If you see a PortAudio error, install it with brew install portaudio (macOS) and reinstall pyaudio.
• Ollama errors?
  • Make sure you have run ollama serve and pulled the llama3.2 model.
  • The app must be able to connect to http://localhost:11434.
• Buttons require a hard click?
  • Enable "Tap to click" in your Mac's Trackpad settings.
• App looks weird or text is missing?
  • Make sure you are running Python 3.8+ and have all dependencies installed.

---

Example Screenshots

(Add screenshots here after running the app!)

---

Credits

• Built with Tkinter, SpeechRecognition, Ollama, and Llama 3.2.

---

Highly Specific Prompt for GPT-4o to Replicate This App

Prompt:

Build a Python desktop app using Tkinter called "Voice-Powered Language Translator" with the following requirements:

1. Language Selection:

   • Provide two dropdown menus at the top: one for "Translate from" and one for "Translate to".
   • Include at least 30 languages (e.g., English, Spanish, French, German, Chinese, Japanese, etc.).
   • Prevent the user from selecting the same language for both dropdowns (show a warning if attempted).

2. Input Methods:

   • Allow the user to either:
     • Click a large, clearly labeled Record button to start recording speech from the microphone, and a Stop button to end recording (no auto-stop; only user stops it).
     • Or type directly into a large, editable Transcript text area.

3. Visual Timer:

   • While recording, display a timer (in seconds) in the status bar, updating every second.

4. Translation:

   • After recording, automatically transcribe the speech and place it in the Transcript box.
   • Provide a large, clearly labeled Translate button that translates the text in the Transcript box (whether typed or transcribed).
   • Show the translation in a large, read-only Translation text area below.

5. UI/UX:

   • Use a modern, light color palette: light backgrounds, dark text, and visually distinct, color-coded status bar (e.g., green for Ready, yellow for Recording, blue for Processing, red for Error, purple for Done).
   • All buttons (Record, Stop, Clear, Translate) should be large, touch-friendly, and have high-contrast text.
   • Add a Clear button to reset both text areas and the status bar.
   • Add clear, beginner-friendly instructions at the top of the window.
   • Make the window and text areas resizable.

6. Status Bar:

   • Show a persistent status bar at the bottom with color-coded feedback: Ready, Recording, Processing, Error, Done.

7. AI Integration:

   • Use the local Ollama Llama 3.2 model for translation by sending a POST request to http://localhost:11434/api/generate with the following JSON:

     {"model": "llama3.2", "prompt": "Translate from [FROM] to [TO]: [TEXT]", "stream": false}

   • Extract the translation from the response field of the returned JSON.
   • Handle connection errors gracefully (show a user-friendly error message if Ollama is not running).

8. Logging:

   • Log each session (transcript, prompt, translation) to a timestamped text file in a logs/ folder.

9. Dependencies:

   • Only require Python, Tkinter, SpeechRecognition, requests, pyaudio, and Ollama (with Llama 3.2 model).

10. Error Handling:

    • Show clear, user-friendly error messages for microphone issues, API errors, or invalid input.
    • Prevent the app from crashing on any user action.

11. Accessibility:

    • Ensure all text is readable (large font, high contrast).
    • All actions should be accessible by mouse or touch (no keyboard required).

12. Beginner-Friendliness:

    • The app should be easy to use for someone with no programming experience.
    • Include clear instructions and feedback for every action.

13. General:

    • The app should be robust, visually appealing, and require no internet connection for translation.
    • The code should be clean, well-commented, and easy to follow.

---