How to Convert a Video into Still Images (Automatically, with URL Examples)

Need to pull thumbnails, preview frames, or scene-specific stills from a video? Whether you’re building a media gallery, generating product snapshots, or just improving UX, converting video to images is a common requirement. But doing it manually is time-consuming. Let’s explore how to automate the process and extract high-quality still images programmatically and efficiently.

Hey everyone,

I’m working on a content site where users upload short videos. I’d like to generate still images from each video for use as preview thumbnails or static placeholders.

So here’s my question: How can I convert a video into still images? Specifically, I’m looking for:

• A way to grab a single frame (e.g., at 2 seconds in)
• Bonus if I can extract multiple frames or use AI to pick a good one
• Ideally, something fast, dynamic, and automatable (not just FFmpeg scripting)

I’m already using Cloudinary for image delivery. Can I do this with their platform?

Yes, and you don’t need to write custom FFmpeg scripts or run a batch job to do it.

With Cloudinary, you can dynamically extract still images from any video file using URL-based transformations. No extra setup, no processing delays, just clean, CDN-delivered images, right when you need them.

Let’s walk through it.

It typically refers to:

• Extracting a single frame as a thumbnail or poster
• Generating a series of frames for galleries or previews
• Capturing keyframes for visual summaries
• Using AI to pick meaningful frames (with face detection)

Cloudinary lets you extract a still frame at a specific timestamp using the so (start offset) transformation.

HTML
<img src="https://res.cloudinary.com/demo/video/upload/so_2/sample.jpg" alt="Video Thumbnail" />Code language: HTML, XML (xml)

• so_2 → Start offset at 2 seconds
• The .jpg extension tells Cloudinary to render it as an image
• Cloudinary automatically picks a high-quality still from that moment

Image from video docs →

HTML
<img src="https://res.cloudinary.com/demo/video/upload/so_2,w_600,h_400,c_fill/sample.jpg" />Code language: HTML, XML (xml)

• w_600,h_400,c_fill → Resize and crop the image
• Use g_auto or g_face to focus on key regions

Need a series of thumbnails? Use the Cloudinary API to request multiple frames from different timestamps.

Example using the Upload API:

Shell
POST /video/upload

{

  "public_id": "product_demo",

  "eager": [

    { "format": "jpg", "transformation": [ { "start_offset": "3" } ] },

    { "format": "jpg", "transformation": [ { "start_offset": "5" } ] },

    { "format": "jpg", "transformation": [ { "start_offset": "10" } ] }

  ]

}Code language: JavaScript (javascript)

Want your stills to center on a face or main subject? Just add g_face or g_auto:

HTML
<img src="https://res.cloudinary.com/demo/video/upload/so_2,c_crop,w_300,h_300,g_face/sample.jpg" />Code language: HTML, XML (xml)

• Cloudinary will scan the frame and crop around the detected face
• Great for profile videos, product closeups, or testimonials

Use Case	How Cloudinary Helps
Product previews	Extract key frame from demo videos
Hero banners	Capture high-res still for header images
Thumbnail grids	Generate stills in consistent size
Social cards	Overlay text on extracted frame
Visual search	Tag extracted stills with AI

Task	Cloudinary Solution
Extract one frame	Add so_2 to the video URL and use .jpg extension
Resize/crop thumbnail	Use w_, h_, c_, g_ parameters
Get multiple frames	Use eager transformations via Upload API
Auto-detect subject	Use g_face or g_auto in crop
Optimize delivery	Delivered via global CDN as optimized image

Cloudinary makes it easy to extract, transform, and deliver still images from video dynamically, without manual editing or heavy scripts.

• 🔗 Video-to-Image Docs
• 🔗 Image Transformation Reference

Generate and customize video thumbnails with just a few clicks. Sign up with Cloudinary to easily enhance the visuals of your videos.