Real-Time Speech Translation Script with Azure & OpenAI

Most commercial real-time speech translation solutions are expensive and complex (about 0.1 usd for 1 min), making them less accessible for students.
This project provides a simple, lightweight, and cost-effective SaaS script
that leverages Azure Speech-to-Text (STT) and OpenAI GPT-4o Mini API for real-time speech translation.

Perfect for students who need real-time subtitles for (online and offline) courses, lectures, or study sessions!
It runs directly in the terminal, displays translations in real time, and automatically logs results to a file.

---

Features

• Real-time Speech Recognition: Uses Azure Speech-to-Text to transcribe speech.
• AI-powered Translation: Uses OpenAI GPT-4o Mini for fast and accurate translation with a avaliable price.
• Terminal Output: Displays both the original and translated text.
• Automatic Logging: Saves transcription and translation results with a timestamp.
• Adjustable Recognition Interval: Modify the recognition interval to balance performance.

---

** How to Use**

Using this script is very simple

1. Clone the Repository

git clone https://github.com/563088655/translate.git
cd translate
pip install azure-cognitiveservices-speech openai

2. Set the key and language

As be wittened at the code, you can chage them at here:

Firistlly, set your API keys and region:

OPENAI_API_KEY = "your-api-key"
AZURE_SPEECH_KEY = "your-azure-speech-key"
AZURE_REGION = "your-area"

And set the speech recognition language (e.g., "ja-JP" for Japanese, "en-US" for English). Check the official Azure documentation for more language options.

speech_config.speech_recognition_language = "set a labguage"

Set the source and target languages (e.g., ja for Japanese, en for English). Check the OpenAI API documentation for supported languages.

def translate_text(text, source_lang="source_language", target_lang="target_lanhguage"):
    if not text:
        return None

3. Just run the script to translate

Open your terminal and run:

python translate_at_terminal.py

Now, you can follow courses in foreign languages with real-time subtitles and automatically save lecture notes for later review!

---

Example log file:

A log file (subtitles_YYYY-MM-DD_HH-MM-SS.txt) is created:

[12:30:10] これはテストです。
[12:30:10] This is a test.

---

Thanks

I read this file to check the openai api's usage, install of the offical page.

https://chatgpt-lab.com/n/nde17253a9bc7

---

Fearure Work

I will develop a simple subtitle software for macOS in the future.
Currently, I am learning SwiftUI and AVFoundation to create a native macOS application
that provides real-time speech recognition, translation, and subtitle overlay.

I am also considering integrating a local AI translator to reduce costs and enhance privacy
（mainly because I'm broke）