StreamShot 📸

Capture high-quality JPEG snapshots from any YouTube livestream via a simple web UI or API. (yes it's a yt-dlp wrapper)

Features

Stream Extraction: Uses yt-dlp to find the best available stream URL.
Frame Capture: Uses OpenCV to grab a single frame directly from the video stream.
Web UI: A clean, Tailwind CSS-powered interface for quick captures.
API Access: Returns snapshots as base64-encoded buffers in a JSON object.

API Usage

Endpoint

POST /snapshot

Request Body

{
  "url": "https://www.youtube.com/watch?v=..."
}

Response

{
  "message": "done",
  "buffer": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAIBAQEBAQIBAQECAgICAg..."
}

CLI Example (Save to File)

To capture a snapshot and save it directly to a file using curl, jq, and base64:

curl -X POST \
  https://shsf-api.reversed.dev/api/exec/16/cca815a2-216d-4362-b6f5-1d57ac9f088d/snapshot \
  -H "Content-Type: application/json" \
  -d '{"url": "YOUR_YOUTUBE_URL_HERE"}' \
  | jq -r .buffer | base64 -d > snapshot.jpg

Local Development

Prerequisites

Python 3.14 (tested)
opencv-python
yt-dlp
numpy

Installation

pip install -r requirements.txt

Running Locally

You can test the capture logic directly:

python main.py

Tech Stack

Backend: Python (yt-dlp, OpenCV)
Frontend: HTML5, Tailwind CSS, JavaScript (Fetch API)
Deployment: Shsf (Serverless)

1.5 KiB Raw Blame History