Pilko Studio is a frame capture web tool that lets you extract screenshots from your videos in many different ways, manually or automated, locally within your browser. It works seamlessly on both desktop and mobile.
Try it here: https://pilko.studio
Offline: Download index.html
Pilko_Demo.mp4
- Example video for first testing.
- Video metadata display (resolution, duration, aspect ratio, estimated video bitrate).
- Capture format (PNG, JPEG, WEBP) and quality configuration (1-100%).
- Configure capture crop area.
- Aspect ratio presets.
- Automatically crop black bars/letterbox.
- Set framerate for frame-accurate captures.
- Capture methods
Manual- Navigate video player to capture and save specific frames.Interval- Extract frames at regular intervals (frames or seconds).Automatic- Automatically detects scene changes and aims to capture only one frame per scene.Transcription- Captures a frame for each timestamp segment.Grid- Generates a grid of screenshots to a single image.
- Add different overlays
- SMPTE timecode
- Custom text
- Upload image/logo
- Configure filenames for export.
- Preview frames in a gallery/lightbox, remove unwanted frames, download individually or as ZIP.
- No cloud processing or installations.
- Frames are stored temporarily in your browser.
- 100% client-side: built into a single
.htmlusing native HTML, CSS and JavaScript.- Why? For trust and accessibility. Easy to deploy locally, no installations required.
thumbnailslecture screenshotsvisual notesdesign referencesmood boardfilm analysisAI dataset buildingMidjourney style references
The tool finds timestamps (or you pick them). Your browser's built-in capabilities (<video>, <canvas>) capture the frame at that time, optionally cropping based on black bar detection, encode it to the selected image format, and provide the image data for display/download.
PySceneDetect exists for better scene detection, but requires Python and setup for local deployment. Pilko Studio's solution is Histogram and Pixel Difference algorithms in Automatic, which get the same job done, are lightweight (basic JS) and work inside a single .html -file. The project's goal is accessibility in deployment for the average users.
The tool doesn't do video extraction from links, such as Youtube, TikTok, Instagram, etc.
I recommend using Media Download or Hitomi Downloader

