### Clone Repository and Run Setup Script

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Clone the TranscriptAI repository and navigate to the CLI directory. Execute the setup script to automatically install Python dependencies, set up a virtual environment, and install FFmpeg.

```bash
git clone https://github.com/ombharatiya/transcript-ai.git
cd transcript-ai/cli
python setup.py
```

--------------------------------

### CLI Installation and Setup

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Clone the repository and run the setup script for the CLI tool. Activate the virtual environment before running transcription commands.

```bash
git clone https://github.com/ombharatiya/transcript-ai.git
cd transcript-ai/cli

python setup.py

source venv/bin/activate
python src/audio_transcriber.py input/audio.mp3
```

--------------------------------

### Manual CLI Setup Alternative

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Manually set up the virtual environment, install Python dependencies, and install FFmpeg if the automated script is not used.

```bash
# Manual setup alternative
python3 -m venv venv
source venv/bin/activate          # Windows: venv\Scripts\activate
pip install -r requirements.txt
brew install ffmpeg               # macOS; see README for Linux/Windows
```

--------------------------------

### FFmpeg Installation for CLI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Instructions for installing FFmpeg, a dependency for audio processing. Use the provided setup script or manual installation methods for macOS, Ubuntu/Debian, or Windows.

```bash
# Install ffmpeg using the setup script:
python setup.py

# Or install manually:
# macOS
brew install ffmpeg

# Ubuntu/Debian  
sudo apt install ffmpeg

# Windows: Download from https://ffmpeg.org/download.html
```

--------------------------------

### Manual CLI Setup - Install FFmpeg

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Instructions for installing FFmpeg on macOS, Ubuntu/Debian, and Windows. FFmpeg is required for audio processing.

```bash
# macOS
brew install ffmpeg
```

```bash
# Ubuntu/Debian
sudo apt update && sudo apt install ffmpeg
```

```bash
# Windows
# Download from https://ffmpeg.org/download.html and add to PATH
```

--------------------------------

### CLI Manual Installation

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Manually set up the Python virtual environment and install dependencies for the CLI tool if the automatic setup is not used.

```bash
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
```

--------------------------------

### Quick Start: Set Up and Transcribe Audio

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/README.md

Navigate to the CLI directory, set up the environment, activate the virtual environment, and then transcribe an audio file.

```bash
# Navigate to CLI directory
cd cli

# Set up environment
python setup.py

# Activate environment  
source venv/bin/activate

# Transcribe audio
python src/audio_transcriber.py input/audio.mp3
```

--------------------------------

### Clone and Run Automated Setup Script

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Use this command to clone the repository and run the automated setup script, which installs dependencies and FFmpeg.

```bash
# Clone and run one-command setup
git clone https://github.com/ombharatiya/transcript-ai.git
cd transcript-ai/cli
python setup.py

# Expected output:
# 🎵 Audio Transcription Tool Setup
# ========================================
# ✓ Python 3.11.x detected
# Creating virtual environment...
# ✓ Virtual environment created
# ✓ Python dependencies installed
# ✓ ffmpeg installed via Homebrew
# ========================================
# 🎉 Setup complete!
# 1. Activate environment: source venv/bin/activate
# 2. Test installation: python src/audio_transcriber.py --help
# 3. Transcribe audio: python src/audio_transcriber.py your_audio.mp3
```

--------------------------------

### Manual CLI Setup - Install Dependencies

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Install the required Python dependencies for the CLI tool using pip after activating the virtual environment.

```bash
pip install -r requirements.txt
```

--------------------------------

### Install and Use TranscriptAI CLI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/docs/index.html

Instructions for installing the TranscriptAI CLI tool and performing basic audio transcription.

```bash
# Install TranscriptAI CLI
git clone https://github.com/ombharatiya/transcript-ai
cd transcript-ai/cli
python setup.py
```

```bash
# Transcribe audio
python src/audio_transcriber.py input/audio.mp3
```

--------------------------------

### Logging Configuration and Examples

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

CLI logging details, including file location, example log entries, and programmatic access to logger settings.

```python
# Log file location
# cli/logs/transcription_20240810.log

# Example log entries:
# 2024-08-10 14:30:01,234 - INFO - Loading Whisper model (medium)...
# 2024-08-10 14:30:08,891 - INFO - Model loaded successfully
# 2024-08-10 14:30:08,892 - INFO - Transcribing: meeting.wav (12.4 MB)
# 2024-08-10 14:31:03,445 - INFO - Text transcription saved to: outputs/meeting_transcription.txt
# 2024-08-10 14:31:03,512 - INFO - Detailed results saved to: outputs/meeting_detailed.json
# 2024-08-10 14:31:03,513 - INFO - Batch transcription completed

# Programmatic log access
import logging
logging.getLogger("audio_transcriber").setLevel(logging.DEBUG)
```

--------------------------------

### Install and Use TranscriptAI CLI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/index.html

Instructions for installing the TranscriptAI CLI tool and performing basic audio transcription and batch processing.

```bash
git clone https://github.com/ombharatiya/transcript-ai
cd transcript-ai/cli
python setup.py
```

```bash
python src/audio_transcriber.py input/audio.mp3
```

```bash
python src/audio_transcriber.py input/*.wav --batch
```

```bash
python src/audio_transcriber.py audio.mp3 \
  --model large \
  --language en \
  --task translate
```

--------------------------------

### Manual CLI Setup - Virtual Environment

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Manually create and activate a Python virtual environment for the CLI tool. This is an alternative to the automatic setup script.

```bash
python3 -m venv venv
source venv/bin/activate
```

```bash
venv\Scripts\activate
```

--------------------------------

### Web Interface - OpenAI API Call (Raw Fetch)

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Example of making a direct POST request to the OpenAI audio transcriptions endpoint using the Fetch API.

```APIDOC
## Web Interface — OpenAI API Call (Raw `fetch`)

### Description
Demonstrates how to directly call the OpenAI transcriptions endpoint from the browser using `fetch` and `FormData`.

### Method
`POST`

### Endpoint
`https://api.openai.com/v1/audio/transcriptions`

### Request Body (FormData)
- **file** (File) - Required - The audio file to transcribe.
- **model** (string) - Required - Set to `whisper-1`.
- **language** (string) - Optional - ISO 639-1 code for the audio language.
- **response_format** (string) - Optional - Set to `verbose_json` for detailed output.

### Request Example
```javascript
const formData = new FormData();
formData.append('file', audioFileObject);
formData.append('model', 'whisper-1');
formData.append('language', 'en');
formData.append('response_format', 'verbose_json');

const response = await fetch('https://api.openai.com/v1/audio/transcriptions', {
    method: 'POST',
    headers: {
        'Authorization': `Bearer sk-...your-key...`
    },
    body: formData
});

if (!response.ok) {
    // Handle errors: 401 (Invalid API key), 429 (Rate limit), 413 (File too large)
}

const data = await response.json();
```

### Response Example (Success)
```json
{
  "text": "Full transcription text",
  "language": "en",
  "duration": 222.4,
  "segments": [
    { "id": 0, "start": 0.0, "end": 4.2, "text": " Hello everyone." },
    ...
  ]
}
```
```

--------------------------------

### Transcribe Audio with Smaller Model

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Use the 'small' model for faster audio processing. Ensure Python 3.8+ is installed.

```bash
python src/audio_transcriber.py file.mp3 --model small
```

--------------------------------

### Create Feature Branch for Web Enhancement

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/README.md

Example Git command to create a new branch for developing a web interface feature. Followed by making changes and submitting a pull request.

```bash
git checkout -b feature/web-enhancement
```

--------------------------------

### Current CLI Project Structure

Source: https://github.com/ombharatiya/transcript-ai/blob/main/PROJECT_STRUCTURE.md

This is the existing directory structure focused on a Python CLI tool. It includes source code, sample input/output, configuration, temporary files, logs, dependencies, and setup scripts.

```tree
transcript-ai/
├── src/                          # CLI source code
├── input/                        # Sample audio files
├── outputs/                      # Sample transcriptions
├── config/                       # Configuration files
├── temp/                         # User temporary files
├── logs/                         # Application logs
├── requirements.txt              # Python dependencies
├── setup.py                      # CLI setup script
└── README.md                     # Main documentation
```

--------------------------------

### Example JSON Output Structure

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

This is an example of the detailed JSON output format, which includes the transcription text, segment information, word-level timestamps, and metadata.

```json
{
  "text": "Your transcribed text here...",
  "segments": [
    {
      "id": 0,
      "seek": 0,
      "start": 0.0,
      "end": 5.0,
      "text": " Your transcribed text here...",
      "tokens": [50364, 2396, ...],
      "temperature": 0.0,
      "avg_logprob": -0.5,
      "compression_ratio": 1.2,
      "no_speech_prob": 0.1
    }
  ],
  "language": "en",
  "metadata": {
    "file_info": {
      "filename": "audio.mp3",
      "size_mb": 5.2,
      "format": ".mp3",
      "supported": true
    },
    "model_used": "base",
    "transcription_time": "0:00:30.123456",
    "timestamp": "2024-01-01T12:00:00.000000"
  }
}
```

--------------------------------

### Get Audio File Metadata

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Retrieves metadata for an audio file, including filename, size, and format, without loading the transcription model. Checks if the format is supported.

```python
transcriber = AudioTranscriber()

info = transcriber.get_audio_info("input/sample_meeting.wav")
print(info)
# {
#   'filename': 'sample_meeting.wav',
#   'size_mb': 12.4,
#   'format': '.wav',
#   'supported': True
# }

info_bad = transcriber.get_audio_info("input/recording.xyz")
print(info_bad)
# {
#   'filename': 'recording.xyz',
#   'size_mb': 3.1,
#   'format': '.xyz',
#   'supported': False
# }
```

--------------------------------

### Set Up Web Development Server

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Steps to set up and run the web development server locally.

```bash
# Set up web development
cd transcript-ai/web
npm install  # if using build tools
python -m http.server 8000  # serve locally
```

--------------------------------

### Clone Repository and Set Up CLI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Instructions for forking the TranscriptAI repository and setting up the CLI for development.

```bash
# Fork the repository
git clone https://github.com/yourusername/transcript-ai.git

# Set up CLI development
cd transcript-ai/cli
python setup.py
```

--------------------------------

### Display Help and Advanced Options

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/README.md

Access comprehensive documentation and available options by running the script with the --help flag.

```bash
python src/audio_transcriber.py --help
```

--------------------------------

### Manual Web Interface Deployment

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/README.md

Commands for manually building and deploying the web interface. This involves copying source files to the 'docs/' directory, committing, and pushing.

```bash
cd web
cp -r src/* ../docs/
git add ../docs/
git commit -m "Update web interface"
git push origin main
```

--------------------------------

### Manual Deployment of TranscriptAI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/DEPLOYMENT_GUIDE.md

Execute these commands to manually build and deploy the web application if automated deployment fails. Ensure you are in the correct directory before running.

```bash
# If automated deployment fails, you can deploy manually
cd web
npm run build
npm run deploy
git add ../docs/
git commit -m "Manual deployment update"
git push origin main
```

--------------------------------

### Serve Web Application Locally with http-server

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/README.md

Alternative command to serve the web application locally using the 'http-server' Node.js package. Ensure you are in the 'web' directory.

```bash
npx http-server src/
```

--------------------------------

### Batch Process Audio Files with TranscriptAI CLI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/docs/index.html

Demonstrates how to use the TranscriptAI CLI for batch processing multiple audio files.

```bash
# Batch processing
python src/audio_transcriber.py input/*.wav --batch
```

--------------------------------

### Advanced TranscriptAI CLI Options

Source: https://github.com/ombharatiya/transcript-ai/blob/main/docs/index.html

Shows how to use advanced options with the TranscriptAI CLI, such as specifying the model, language, and task.

```bash
# Advanced options
python src/audio_transcriber.py audio.mp3 \
  --model large \
  --language en \
  --task translate
```

--------------------------------

### Serve Web Application Locally with Python

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/README.md

Command to serve the web application locally using Python's built-in HTTP server. Ensure you are in the 'web' directory.

```bash
python -m http.server 8000
```

--------------------------------

### Initialize AudioTranscriber with Default Settings

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Initializes the AudioTranscriber class using default settings for model size ('base') and output directory ('./outputs/').

```python
from src.audio_transcriber import AudioTranscriber

# Default: 'base' model, output to ./outputs/
transcriber = AudioTranscriber()
```

--------------------------------

### Check Supported File Formats

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Verify supported audio file formats using the --info flag. For unsupported formats, FFmpeg can be used for conversion.

```bash
python src/audio_transcriber.py input/file.xyz --info
```

```bash
ffmpeg -i input/file.xyz -c:a aac input/output.aac
```

--------------------------------

### File Information CLI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Use this command to retrieve details about an audio file before transcription.

```bash
# Check audio file details
python src/audio_transcriber.py input/audio.mp3 --info
```

--------------------------------

### Basic CLI Transcription

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Use this command to transcribe a single audio file. Ensure you are in the CLI directory and have activated your virtual environment.

```bash
# Navigate to CLI directory and activate environment
cd transcript-ai/cli
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Transcribe a single file
python src/audio_transcriber.py input/audio.mp3

# Use a different model
python src/audio_transcriber.py input/audio.wav --model large

# Specify language
python src/audio_transcriber.py input/audio.aac --language en
```

--------------------------------

### AudioTranscriber Initialization

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Initializes the Whisper model wrapper, creates the output directory, configures log files, and auto-detects FFmpeg. Supports specifying model size and output directory.

```APIDOC
## AudioTranscriber Initialization

`AudioTranscriber(model_size, output_dir)` initializes the Whisper model wrapper, creates the output directory, configures daily rotating log files, and auto-detects a local or system FFmpeg binary.

```python
from src.audio_transcriber import AudioTranscriber

# Default: 'base' model, output to ./outputs/
transcriber = AudioTranscriber()

# High-accuracy setup with custom output directory
transcriber = AudioTranscriber(model_size="large-v3", output_dir="/data/transcripts")

# Available model sizes (speed ↔ accuracy trade-off):
# tiny (39MB) | base (74MB) | small (244MB) | medium (769MB)
# large (1.5GB) | large-v2 (1.5GB) | large-v3 (1.5GB)
```
```

--------------------------------

### Clone TranscriptAI Repository

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/README.md

Command to clone the TranscriptAI repository from GitHub. This is the first step for local development.

```bash
git clone https://github.com/ombharatiya/transcript-ai.git
cd transcript-ai/web
```

--------------------------------

### CLI for Meeting Transcription

Source: https://github.com/ombharatiya/transcript-ai/blob/main/shared/examples/USE_CASES.md

Use the command-line interface to transcribe meeting audio files. Specify the model for desired accuracy and speed.

```bash
python src/audio_transcriber.py meetings/weekly-standup.mp3 --model medium
```

--------------------------------

### Default Configuration JSON

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

The default configuration file for the CLI, specifying model, language, output, and batch settings.

```json
{
  "model_settings": {
    "default_model": "base",
    "available_models": ["tiny", "base", "small", "medium", "large", "large-v2", "large-v3"],
    "default_language": null,
    "default_task": "transcribe",
    "default_temperature": 0.0
  },
  "output_settings": {
    "output_directory": "outputs",
    "save_json_details": true,
    "log_directory": "logs",
    "timestamp_filenames": false
  },
  "supported_formats": [".aac", ".flac", ".mp3", ".mp4", ".m4a", ".ogg", ".wav", ".webm"],
  "batch_settings": {
    "max_concurrent": 1,
    "continue_on_error": true
  }
}
```

--------------------------------

### GitHub Actions Workflow for Web Deployment

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/README.md

Configuration snippet for a GitHub Actions workflow that automates the deployment of the web interface to GitHub Pages. It copies source files to the 'docs/' directory for serving.

```yaml
# .github/workflows/deploy-web.yml
- Web source files are copied from web/src/ to docs/
- GitHub Pages serves from docs/ directory  
- Live at: https://ombharatiya.github.io/transcript-ai
```

--------------------------------

### Initialize AudioTranscriber with Custom Settings

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Initializes the AudioTranscriber class with a specified model size ('large-v3') and a custom output directory for transcripts.

```python
# High-accuracy setup with custom output directory
transcriber = AudioTranscriber(model_size="large-v3", output_dir="/data/transcripts")
```

--------------------------------

### CLI - Command Reference

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Provides command-line interface instructions for using the Transcript AI tool, covering single file transcription, batch processing, translation, and file information retrieval.

```APIDOC
## CLI — Command Reference

The `main()` function in `audio_transcriber.py` exposes all `AudioTranscriber` capabilities as command-line flags.

### Usage

1.  **Activate virtual environment:**
    ```bash
    source venv/bin/activate
    ```

2.  **Single File Transcription:**
    ```bash
    python src/audio_transcriber.py <audio_file_path>
    ```
    *Example:* `python src/audio_transcriber.py input/audio.mp3`

3.  **Single File with Options:**
    ```bash
    python src/audio_transcriber.py <audio_file_path> --model <model_size> --language <lang_code> --output-dir <output_directory>
    ```
    *Example:* `python src/audio_transcriber.py input/audio.wav --model large-v3 --language en --output-dir /data/results`

4.  **Translation:**
    ```bash
    python src/audio_transcriber.py <audio_file_path> --task translate --language <source_lang_code> --temperature <temp_value>
    ```
    *Example:* `python src/audio_transcriber.py input/french_lecture.mp3 --task translate --language fr --temperature 0.2`

5.  **Batch Mode Transcription:**
    ```bash
    python src/audio_transcriber.py <input_files_pattern> --batch --model <model_size>
    ```
    *Example:* `python src/audio_transcriber.py input/*.mp3 --batch --model medium`

6.  **Batch Mode with Multiple Explicit Files:**
    ```bash
    python src/audio_transcriber.py <file1> <file2> ... --model <model_size> --output-dir <output_directory>
    ```
    *Example:* `python src/audio_transcriber.py file1.wav file2.mp3 file3.m4a --model small --output-dir outputs/project`

7.  **File Information (No Transcription):**
    ```bash
    python src/audio_transcriber.py <audio_file_path> --info
    ```
    *Example:* `python src/audio_transcriber.py input/audio.mp3 --info`
    *Output Example:* `AUDIO FILE INFORMATION`
```

--------------------------------

### Podcast Show Notes Generation

Source: https://github.com/ombharatiya/transcript-ai/blob/main/shared/examples/USE_CASES.md

Generate show notes for a series of podcast episodes using batch processing. Specify an output directory for the generated files.

```bash
python src/audio_transcriber.py podcasts/season1/*.mp3 --batch --model large --output-dir show-notes/
```

--------------------------------

### Batch Processing with CLI

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Process multiple audio files efficiently using batch mode. You can specify file patterns or list individual files, and customize model and language settings.

```bash
# Process multiple files
python src/audio_transcriber.py input/*.mp3 --batch

# Process specific files
python src/audio_transcriber.py input/file1.wav input/file2.aac input/file3.mp3

# Batch with custom settings
python src/audio_transcriber.py input/*.wav --batch --model medium --language es
```

--------------------------------

### Run CLI for File Information Only

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Obtain file information without performing transcription using the --info flag.

```python
python src/audio_transcriber.py audio.mp3 --info
```

--------------------------------

### GitHub Actions CI/CD Workflows

Source: https://github.com/ombharatiya/transcript-ai/blob/main/PROJECT_STRUCTURE.md

This section outlines the CI/CD workflows for the monorepo. It includes configurations for testing the Python CLI, building the web interface, and deploying to GitHub Pages.

```yaml
# CI for Python CLI
# Build web interface
# Deploy to GitHub Pages
```

--------------------------------

### Run Audio Transcription via CLI

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Execute audio transcription using the command-line interface. Specify input files and options like model, language, and output format.

```bash
python src/audio_transcriber.py input/audio.mp3 --no-json
```

```bash
python src/audio_transcriber.py --help
```

--------------------------------

### Run CLI and Adjust Temperature

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Control the randomness of the transcription output by adjusting the temperature parameter.

```python
python src/audio_transcriber.py audio.mp3 --temperature 0.2
```

--------------------------------

### Transcription with Different Model Sizes

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/README.md

Specify the desired AI model size (e.g., 'large') to balance speed and accuracy for transcription.

```bash
# Different model sizes
python src/audio_transcriber.py input/audio.wav --model large
```

--------------------------------

### CLI for Audio Transcription

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

The command-line interface allows transcription of single files, translation, batch processing, and retrieving file information without transcription. Use flags like `--model`, `--language`, `--task`, `--batch`, and `--info` to control behavior.

```bash
# Activate virtualenv first
source venv/bin/activate

# --- Single file ---
python src/audio_transcriber.py input/audio.mp3

# With model and language
python src/audio_transcriber.py input/audio.wav \
  --model large-v3 \
  --language en \
  --output-dir /data/results

# --- Translation ---
python src/audio_transcriber.py input/french_lecture.mp3 \
  --task translate \
  --language fr \
  --temperature 0.2

# --- Batch mode ---
python src/audio_transcriber.py input/*.mp3 --batch --model medium

# Multiple explicit files
python src/audio_transcriber.py file1.wav file2.mp3 file3.m4a \
  --model small \
  --output-dir outputs/project

# --- File info (no transcription) ---
python src/audio_transcriber.py input/audio.mp3 --info
# ============================================================
# AUDIO FILE INFORMATION

```

--------------------------------

### Run CLI with Custom Output Directory

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Specify a custom directory for output files when using the CLI tool.

```python
python src/audio_transcriber.py audio.mp3 --output-dir results/
```

--------------------------------

### Real-time Transcription for Accessibility

Source: https://github.com/ombharatiya/transcript-ai/blob/main/shared/examples/USE_CASES.md

Transcribe live audio streams for accessibility purposes. The 'auto' language detection is useful for unknown audio sources.

```bash
python src/audio_transcriber.py live-meeting.wav --model base --language auto
```

--------------------------------

### CLI Usage

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Command-line interface for audio transcription. Supports various models, languages, and output options.

```APIDOC
## CLI Usage

### Description
Command-line interface for audio transcription. Supports various models, languages, and output options.

### Usage
```bash
python src/audio_transcriber.py input/audio.mp3 --no-json
```

### Options Reference

#### Positional Arguments
- **files** (string) - One or more audio files

#### Optional Arguments
- **--model** (string) - Specifies the transcription model. Options: tiny, base, small, medium, large, large-v2, large-v3. Default: base.
- **--language** (string) - ISO 639-1 code for the audio language (e.g., en, es, fr). Default: auto-detect.
- **--task** (string) - Task to perform: transcribe or translate. Default: transcribe.
- **--temperature** (float) - Controls randomness. Range: 0.0 (deterministic) to 1.0 (creative). Default: 0.0.
- **--output-dir** (string) - Directory to save output files. Default: outputs.
- **--no-json** (boolean) - Omit detailed JSON output.
- **--batch** (boolean) - Enable explicit batch mode.
- **--info** (boolean) - Print file metadata and skip transcription.
```

--------------------------------

### Basic Transcription: Single Audio File

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/README.md

Use this command to transcribe a single audio file. Ensure the input file path is correct.

```bash
# Single file transcription
python src/audio_transcriber.py input/meeting.wav
```

--------------------------------

### CLI Basic Usage

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Execute transcriptions using the CLI tool for single files or batch processing. Options are available to specify models, languages, and tasks like translation.

```bash
python src/audio_transcriber.py input/audio.mp3
```

```bash
python src/audio_transcriber.py input/*.wav --batch
```

```bash
python src/audio_transcriber.py audio.mp3 --model large
```

```bash
python src/audio_transcriber.py audio.mp3 --language en
```

```bash
python src/audio_transcriber.py foreign.mp3 --task translate
```

--------------------------------

### Batch Lecture Transcription

Source: https://github.com/ombharatiya/transcript-ai/blob/main/shared/examples/USE_CASES.md

Process an entire semester's worth of lecture recordings in batch mode. Larger models provide higher accuracy.

```bash
python src/audio_transcriber.py lectures/*.mp3 --batch --model large
```

--------------------------------

### Transcribe Multiple Audio Files in Batch

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/input/README.md

This command transcribes all audio files matching the specified pattern (e.g., *.wav) in the 'input' directory. The --batch flag enables this functionality.

```bash
python src/audio_transcriber.py input/*.wav --batch
```

--------------------------------

### Transcribe Single Audio File

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/input/README.md

Use this command to transcribe a single audio file. Ensure the file is placed in the 'input' directory.

```bash
python src/audio_transcriber.py input/meeting.mp3
```

--------------------------------

### Advanced CLI Transcription Options

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

These commands demonstrate advanced features like translation, adjusting transcription temperature, specifying output directories, and skipping JSON details.

```bash
# Translate to English
python src/audio_transcriber.py input/foreign_audio.mp3 --task translate

# Adjust sampling temperature for creative/technical content
python src/audio_transcriber.py input/audio.wav --temperature 0.2

# Custom output directory
python src/audio_transcriber.py input/audio.mp3 --output-dir /custom/path

# Skip detailed JSON output
python src/audio_transcriber.py input/audio.mp3 --no-json
```

--------------------------------

### Transcribe Audio File with Custom Output Directory

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/input/README.md

Specify a custom output directory for transcription results using the --output-dir flag. This is useful for organizing your transcripts.

```bash
python src/audio_transcriber.py input/podcast.m4a --output-dir transcripts/
```

--------------------------------

### Reduce Model Size for Memory Errors

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

If you encounter out-of-memory errors, use a smaller Whisper model like 'tiny' to reduce resource consumption.

```bash
# Use a smaller model
python src/audio_transcriber.py large_file.mp3 --model tiny
```

--------------------------------

### AudioTranscriber.batch_transcribe()

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Processes a list of audio files sequentially. It loads the model only once for efficiency, captures per-file errors without stopping the batch, and returns a dictionary where keys are file paths and values are the results or error messages.

```APIDOC
## `AudioTranscriber.batch_transcribe()` — Multiple Files

Processes a list of audio files sequentially, loads the model once, captures per-file errors without aborting the batch, and returns a dict keyed by file path.

### Method Signature
`batch_transcribe(audio_paths: List[str], language: Optional[str] = None, task: str = "transcribe", temperature: float = 0.0, save_json: bool = False) -> Dict[str, Dict]`

### Parameters
- **audio_paths** (List[str]) - A list of paths to the audio files to process.
- **language** (Optional[str]) - The language of the audio files. If `None`, the language will be auto-detected for each file. Defaults to `None`.
- **task** (str) - The task to perform, either "transcribe" (default) or "translate". Defaults to "transcribe".
- **temperature** (float) - Controls the randomness of the predictions. Lower values make the output more deterministic. Defaults to `0.0`.
- **save_json** (bool) - Whether to save the detailed transcription as a JSON file for each audio file. Defaults to `False`.

### Returns
A dictionary where keys are the audio file paths and values are dictionaries containing either the transcription results or an 'error' key with the error message.

### Request Example
```python
import glob
from transcript_ai import AudioTranscriber

transcriber = AudioTranscriber(model_size="small", output_dir="outputs/batch")

# Collect audio files
files = glob.glob("input/*.mp3") + glob.glob("input/*.wav")

results = transcriber.batch_transcribe(
    files,
    language=None,        # Auto-detect per file
    task="transcribe",
    temperature=0.0,
    save_json=True
)

# Process results
successful = {k: v for k, v in results.items() if "error" not in v}
failed     = {k: v for k, v in results.items() if "error" in v}

print(f"Processed: {len(results)} | OK: {len(successful)} | Failed: {len(failed)}")

# Access individual result
print(results["input/sample_meeting.wav"]["text"][:100])

# Inspect errors
for path, res in failed.items():
    print(f"FAILED {path}: {res['error']}")
```
```

--------------------------------

### AudioTranscriber.load_model()

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Loads the Whisper model into memory on the first call. Models are downloaded and cached locally on first use. Subsequent calls are no-ops.

```APIDOC
## `AudioTranscriber.load_model()` — Lazy Model Loading

Loads the Whisper model into memory on first call; subsequent calls are no-ops. Models are downloaded from the internet on first use and cached locally.

```python
transcriber = AudioTranscriber(model_size="medium")

# Explicitly pre-load (optional — transcribe_audio calls this automatically)
transcriber.load_model()
# Logs: "Loading Whisper model (medium)..."
# Logs: "Model loaded successfully"

# Model is now cached: transcriber.model is set
print(transcriber.model)  # <whisper.model.Whisper object>
```
```

--------------------------------

### Run CLI and Skip JSON Output

Source: https://github.com/ombharatiya/transcript-ai/blob/main/web/src/docs.html

Use the --no-json flag to prevent the CLI tool from generating JSON output.

```python
python src/audio_transcriber.py audio.mp3 --no-json
```

--------------------------------

### Transcription with Language Specification

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/README.md

Explicitly set the language of the audio file (e.g., 'en' for English) if automatic detection is not desired or accurate.

```bash
# Language specification
python src/audio_transcriber.py input/audio.aac --language en
```

--------------------------------

### Web Interface OpenAI API Call using fetch

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Making a direct POST request to the OpenAI transcriptions endpoint using FormData and fetch in the browser.

```javascript
const formData = new FormData();
formData.append('file', audioFileObject);   // File object from input/drop
formData.append('model', 'whisper-1');
formData.append('language', 'en');          // Optional ISO 639-1 code
formData.append('response_format', 'verbose_json');

const response = await fetch('https://api.openai.com/v1/audio/transcriptions', {
    method: 'POST',
    headers: {
        'Authorization': `Bearer sk-...your-key...`
    },
    body: formData
});

if (!response.ok) {
    // 401 → Invalid API key
    // 429 → Rate limit exceeded
    // 413 → File too large (>25 MB)
}

const data = await response.json();
// {
//   "text": "Full transcription text",
//   "language": "en",
//   "duration": 222.4,
//   "segments": [
//     { "id": 0, "start": 0.0, "end": 4.2, "text": " Hello everyone." },
//     ...
//   ]
// }
```

--------------------------------

### Podcast Episodes Transcription

Source: https://github.com/ombharatiya/transcript-ai/blob/main/README.md

Transcribe podcast episodes in batch mode using the 'large' model.

```bash
python src/audio_transcriber.py input/podcast_*.mp3 --batch --model large
```

--------------------------------

### Translation Task: Foreign Language to English

Source: https://github.com/ombharatiya/transcript-ai/blob/main/cli/README.md

Use the --task translate flag to transcribe audio in a foreign language and translate it to English.

```bash
# Translation to English
python src/audio_transcriber.py input/foreign.mp3 --task translate
```

--------------------------------

### AudioTranscriber.get_audio_info()

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Returns metadata about an audio file, including filename, size, and format, without loading the transcription model.

```APIDOC
## `AudioTranscriber.get_audio_info()` — File Metadata

Returns a dictionary with filename, file size in MB, format extension, and whether the format is supported — without loading the model or processing audio.

```python
transcriber = AudioTranscriber()

info = transcriber.get_audio_info("input/sample_meeting.wav")
print(info)
# {
#   'filename': 'sample_meeting.wav',
#   'size_mb': 12.4,
#   'format': '.wav',
#   'supported': True
# }

info_bad = transcriber.get_audio_info("input/recording.xyz")
print(info_bad)
# {
#   'filename': 'recording.xyz',
#   'size_mb': 3.1,
#   'format': '.xyz',
#   'supported': False
# }
```
```

--------------------------------

### AudioTranscriber.is_supported_format()

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Checks if a given file extension is among the supported audio formats for transcription.

```APIDOC
## `AudioTranscriber.is_supported_format()` — Format Validation

Checks whether a file's extension belongs to the supported set: `.aac`, `.flac`, `.mp3`, `.mp4`, `.m4a`, `.ogg`, `.wav`, `.webm`.

```python
transcriber = AudioTranscriber()

print(transcriber.is_supported_format("audio.mp3"))   # True
print(transcriber.is_supported_format("audio.flac"))  # True
print(transcriber.is_supported_format("audio.wma"))   # False
print(transcriber.is_supported_format("video.mkv"))   # False

# Convert an unsupported format first, then transcribe
import subprocess
subprocess.run(["ffmpeg", "-i", "input/recording.wma", "-c:a", "aac", "input/recording.aac"])
result = transcriber.transcribe_audio("input/recording.aac")
```
```

--------------------------------

### Trigger GitHub Pages Deployment

Source: https://github.com/ombharatiya/transcript-ai/blob/main/DEPLOYMENT_GUIDE.md

Commit and push changes to the main branch to trigger the GitHub Actions workflow for deployment.

```bash
git add .
git commit -m "Enable GitHub Pages with GitHub Actions workflow"
git push origin main
```

--------------------------------

### Proposed Monorepo Project Structure

Source: https://github.com/ombharatiya/transcript-ai/blob/main/PROJECT_STRUCTURE.md

The proposed monorepo structure organizes the project into distinct components: a Python CLI, a web interface, shared resources, and GitHub Actions workflows. This facilitates unified development and deployment.

```tree
transcript-ai/
├── cli/                          # Python CLI Tool
│   ├── src/
│   │   └── audio_transcriber.py
│   ├── input/                    # Sample audio files
│   ├── outputs/                  # Sample transcriptions  
│   ├── config/
│   ├── requirements.txt
│   ├── setup.py
│   └── README_CLI.md
│
├── web/                          # Web Interface
│   ├── src/
│   │   ├── index.html
│   │   ├── app.js
│   │   └── styles.css
│   ├── public/
│   ├── dist/                     # Built files for GitHub Pages
│   ├── package.json
│   └── README_WEB.md
│
├── docs/                         # GitHub Pages (symbolic link to web/dist)
│
├── shared/                       # Common resources
│   ├── examples/
│   ├── assets/
│   └── docs/
│
├── .github/
│   ├── workflows/
│   │   ├── test-cli.yml         # CI for Python CLI
│   │   ├── build-web.yml        # Build web interface
│   │   └── deploy-pages.yml     # Deploy to GitHub Pages
│   └── ISSUE_TEMPLATE/
│
├── README.md                     # Main project documentation
├── CONTRIBUTING.md
├── LICENSE
└── CHANGELOG.md
```

--------------------------------

### Batch Transcribe Multiple Audio Files

Source: https://context7.com/ombharatiya/transcript-ai/llms.txt

Use `batch_transcribe` to process a list of audio files efficiently. It loads the model only once and handles individual file errors without stopping the entire batch. The results are returned as a dictionary keyed by file path.

```python
import glob
from src.audio_transcriber import AudioTranscriber

transcriber = AudioTranscriber(model_size="small", output_dir="outputs/batch")

# Collect files
files = glob.glob("input/*.mp3") + glob.glob("input/*.wav")

results = transcriber.batch_transcribe(
    files,
    language=None,        # Auto-detect per file
    task="transcribe",
    temperature=0.0,
    save_json=True
)

# Summary
successful = {k: v for k, v in results.items() if "error" not in v}
failed     = {k: v for k, v in results.items() if "error" in v}

print(f"Processed: {len(results)} | OK: {len(successful)} | Failed: {len(failed)}")
# Processed: 8 | OK: 7 | Failed: 1

# Access individual result
print(results["input/sample_meeting.wav"]["text"][:100])
# "Good morning. Let's get started with today's agenda..."

# Inspect errors
for path, res in failed.items():
    print(f"FAILED {path}: {res['error']}")
# FAILED input/corrupt.mp3: Audio file not found: input/corrupt.mp3

```