Parakeet-API: High-Performance Local Speech-to-Text Service

The parakeet-api project is a local speech-to-text service based on the NVIDIA Parakeet-tdt-0.6b model. It provides an OpenAI API-compatible interface and a clean Web user interface, enabling you to easily and quickly convert any audio or video file into high-precision SRT subtitles. It's also compatible with pyVideoTrans v3.72+.

This project's open-source address: https://github.com/jianchang512/parakeet-api

✨ Parakeet-API Key Advantages

• 🚀 Extreme Speed and Performance: The Parakeet model is highly optimized, especially in environments equipped with NVIDIA GPUs, offering very fast transcription speeds, making it ideal for processing large or long audio and video files.
• 🎯 Precise Timestamps: Adopting advanced Transducer (TDT) technology, the generated SRT subtitles have very accurate timestamps, perfectly aligned with the audio stream, ideal for video subtitling.
• 💰 Completely Free, Unlimited Use: Run on your own hardware without paying any API call fees or usage time restrictions.
• 🌐 Flexible Access Methods: Provides an intuitive Web interface and standardized API interface, easily integrated into existing workflows like pyVideoTrans.

🛠️ Installation and Configuration Guide

This project supports Windows, macOS, and Linux. Please follow the steps below for installation and configuration.

Step 0: Configure Python 3.10 Environment

If you don't have Python 3 installed, please follow this tutorial: https://pvt9.com/_posts/pythoninstall

Step 1: Prepare FFmpeg

This project uses ffmpeg for audio and video format preprocessing.

• Windows (Recommended):

  1. Download from the FFmpeg github repository and extract to get ffmpeg.exe.
  2. Place the downloaded ffmpeg.exe file directly in the root directory of this project (at the same level as the app.py file). The program will automatically detect and use it without configuring environment variables.

• macOS (Using Homebrew):

  brew install ffmpeg

• Linux (Debian/Ubuntu):

  sudo apt update && sudo apt install ffmpeg

Step 2: Create a Python Virtual Environment and Install Dependencies

1. Download or clone this project code to your local computer (it's recommended to place it in a non-system drive, inside a folder named with English characters or numbers).

2. Open a terminal or command-line tool and navigate to the project root directory (on Windows, you can directly type cmd in the folder's address bar and press Enter).

3. Create a virtual environment: python -m venv venv

4. Activate the virtual environment:

   • Windows (CMD/PowerShell): .\venv\Scripts\activate
   • macOS / Linux (Bash/Zsh): source venv/bin/activate

5. Install dependencies:

   • If you do not have an NVIDIA graphics card (use CPU only):

     pip install -r requirements.txt

   • If you have an NVIDIA graphics card (using GPU acceleration): a. Ensure you have installed the latest NVIDIA drivers and the corresponding CUDA Toolkit. b. Uninstall any existing old versions of PyTorch: pip uninstall -y torch c. Install PyTorch that matches your CUDA version (using CUDA 12.6 as an example):

     pip install torch --index-url https://download.pytorch.org/whl/cu126

Step 3: Start the Service

In the terminal where the virtual environment is activated, run the following command:

python app.py

You will see a message indicating that the service has started. The first run will download the model (approximately 1.2GB), please be patient.

If a bunch of warnings appear, don't worry about it.

Successful startup interface

🚀 Usage Methods

Method 1: Using the Web Interface

1. Open in your browser: http://127.0.0.1:5092
2. Drag and drop or click to upload your audio or video file.
3. Click "Start Transcription" and wait for the process to complete. You can then see and download the SRT subtitles below.

Method 2: API Call (Python Example)

You can easily call this service using the openai library.

from openai import OpenAI

client = OpenAI(
    base_url="http://127.0.0.1:5092/v1",
    api_key="any-key",
)

with open("your_audio.mp3", "rb") as audio_file:
    srt_result = client.audio.transcriptions.create(
        model="parakeet",
        file=audio_file,
        response_format="srt"
    )
print(srt_result)

Method 3: Integration with pyVideoTrans (Recommended)

Parakeet-API can be seamlessly integrated with the video translation tool pyVideoTrans (v3.72 and above).

1. Ensure your parakeet-api service is running locally.
2. Open the pyVideoTrans software.
3. In the menu bar, select Speech Recognition(R) -> Nvidia parakeet-tdt.
4. In the configuration window that pops up, set the "http address" to: http://127.0.0.1:5092/v1
5. Click "Save" to start using.