YouTube Subtitle Generator

An Apify actor that automatically transcribes YouTube videos and translates the subtitles using OpenAI's AI models. It's designed for on-demand use, not automated workflows, due to YouTube's cookie restrictions.

Overview

This actor takes a YouTube video URL, uses OpenAI's Whisper model for speech recognition, and can translate the transcript into over 15 languages using GPT-4o. The primary output is subtitle files (SRT) and structured text data. A key limitation is that YouTube cookies expire within minutes, so each run requires fresh cookies, making it unsuitable for scheduled tasks.

Key Features

• Automatic Transcription: Uses OpenAI's AI for speech-to-text.
• Translation: Translates transcripts to multiple languages using GPT-4o.
• Batch Processing: Generates a result every 2 minutes of video for efficient processing.
• Multiple Output Formats: SRT subtitle files, plain text (TXT), and structured JSON data.
• Video Metadata: Extracts full video info like title, channel, and view count.
• Language Support: Auto-detects source language. Supports translation to English, Traditional/Simplified Chinese, Japanese, Korean, Spanish, French, German, Italian, Portuguese, Russian, Arabic, Hindi, Thai, and Vietnamese.

How to Use

1. Get YouTube Cookies (Required for Every Run)

Cookies expire quickly and must be manually provided each time.
1. Install the Chrome extension Get cookies.txt LOCALLY.
2. Log into YouTube in your browser and navigate to a video page.
3. Click the extension icon and select "Export as Netscape".
4. Copy the entire text content. This is your youtubeCookies input.

2. Run the Actor

1. Go to the actor on Apify Console (search for "YouTube Subtitle Generator").
2. Configure the input:
   • videoUrl: The YouTube video URL.
   • openaiApiKey: Your valid OpenAI API key.
   • youtubeCookies: The freshly exported Netscape-format cookies.
   • (Optional) Set sourceLanguage, targetLanguage, or gptModel.
3. Click "Start".

Input Parameters

Required

Optional

Output

The actor produces results in batches (every ~2 minutes of video) and a final dataset.

Final outputs include:
* SRT files: Subtitles in both the original and translated languages.
* TXT file: Plain text transcript.
* JSON: Structured data containing the full transcript segments, translations, and video metadata.
* Metadata: Complete video information (title, channel, views, etc.).
* Statistics: Processing stats.

Important Notes & Troubleshooting

️ Critical: YouTube Cookie Limitation

• Cookies expire in minutes. You need new ones for every execution.
• This actor is not suitable for automation or scheduling. It's for manual, on-demand use.
• Always use the Netscape format (text starting with # Netscape HTTP Cookie File).

Best Practices

• Test with shorter videos (under 30 minutes is recommended).
• Ensure a stable network connection.
• Check your OpenAI API quota before processing long videos.

Common Issues

• Video won't download? Confirm your youtubeCookies are fresh and the video is publicly available.
• No transcription output? Verify your openaiApiKey is valid and has not reached its usage limits.
• Translation not working? Ensure sourceLanguage and targetLanguage are different and you have access to the gpt-4o model.

Pricing

Apify charges based on the number of results in the dataset. For example:
* 2-minute video: ~2 results
* 10-minute video: ~6 results
* 30-minute video: ~16 results