VTT 2.0

Extract speech from a video (or audio) file and generate real, timestamped WebVTT/SRT subtitles — or use the VTT Chrome Extension for live online Persian (FA) & English (EN) subtitles on YouTube and any video playing in Chrome.

---

🧩 New: VTT Chrome Extension (vtt-extention)

Minimal online subtitle overlay for YouTube and any HTML5 video playing in Chrome, built for Persian (فارسی) and English (EN):

• Dedicated Repository: vtt-extention
• Works on YouTube & Any Online Video: Automatically attaches to <video> elements in Chrome.
• Persian (FA), English (EN), & Bilingual (FA + EN): Instant language switching from the popup or the subtle top-right video corner chip.
• Calm, Sentence-Paced Reading: Automatically merges rapid word-by-word auto-captions into full, readable 3.5s–5.5s sentence blocks.
• Live Audio Speech Mode (🎤): Generates live online subtitles even for videos without uploaded caption tracks.
• 1-Click .VTT & .SRT Export: Download captured subtitles anytime.

⬇ Free Installation

• Option A — Full Chrome Extension (vtt-chrome-extension.zip):
  1. Download vtt-chrome-extension.zip and extract it to a folder.
  2. Open chrome://extensions/ and turn Developer mode ON.
  3. Drag the extracted folder onto chrome://extensions/ (or click Load unpacked).
• Option B — 1-Click Userscript (vtt-subtitles.user.js):
  • If you have Tampermonkey or Violentmonkey installed, simply click Install vtt-subtitles.user.js for a 1-click install with no unzipping.

---

Release notes (Awakens 🐉)

VTT is finally awake after a long sleep — and it finally makes actual .vtt files. 😄

✨ Highlights

• 🧩 Chrome Extension: Minimal online Persian & English subtitle overlay for YouTube and any video in Chrome (chrome-extension/).
• 🧠 New brain: swapped the old online-only Google Web Speech API in the CLI for Whisper (via faster-whisper). Fully offline, ~100 languages, and much more accurate — especially for Persian/non-English audio.
• 🎬 Real subtitles at last: output is now genuine timestamped .vtt / .srt, not a single blob of plain text.
• 📁 Batch mode: point it at a folder, get subtitles for every video inside.
• ▶️ YouTube/URL input: vtt_run "https://youtube.com/watch?v=..." just works (needs the optional [youtube] extra).
• 🖥️ GPU support: --device cuda for faster runs.
• 🎚 Model size control: --model tiny|base|small|medium|large-v3 to trade off speed vs. accuracy.

⚠️ Breaking change

Default CLI output is now .vtt instead of .txt. Pass --format txt if you relied on the old plain-text output.

📦 Install CLI

pip install video-to-text-vtt
pip install "video-to-text-vtt[youtube]"   # optional, for URL input

🔧 CLI Example

vtt_run my_video.mp4 -lang fa --model medium
vtt_run ./my_videos --format all -o ./subtitles

Migrating from v1

The CLI is backward-compatible for the common case (python main.py path/to/video), but the default output is now a real .vtt file instead of a .txt file. Pass --format txt if you want the old plain-text behavior.