- Get Started
- Image & Video APIs overview
- AI Power Start
- SDK Quick Starts
- Developer Kickstart
- AI agents
- Try it!
- VS Code Extension
- Video tutorial library
- Programmatic asset management
- Image onboarding
- Programmatic upload
- Upload programmatically
- Upload with Fetch API
- Create upload presets (Node.js)
- Create upload presets (Console UI)
- Auto upload
- Moderate images with AI
- Use AI to generate image captions
- Upload images in Flutter
- Upload images in Node.js
- Auto-tag images in Node.js
- Upload multiple files in Node.js
- Upload videos in Node.js
- Upload images in Python
- Auto-tag images in Python
- Upload videos in Python
- Upload images in PHP
- Upload images in Go
- Upload assets in a React app
- Upload assets in a Vue.js app
- Drag-and-drop uploads in React
- Upload assets in a Next.js app
- Upload assets with Server Actions
- Upload assets in Svelte
- Upload assets in a SvelteKit app
- Upload assets in a Remix app
- Upload images in Astro
- Upload images in Hono
- Use webhooks to remove backgrounds
- Upload images in Laravel
- Interaction with Cloudinary APIs
- Cloudinary SDKs
- Find your credentials
- Configure the JavaScript SDK
- Configure the React SDK
- Configure the Angular SDK
- Configure the Flutter SDK
- Configure Svelte Cloudinary
- Configure the Javascript SDK in Svelte
- Getting started with Cloudinary in Node.js
- Configure the Node.js SDK
- Configure the Python SDK
- Configure the PHP SDK
- Install and configure Laravel SDK
- Configure the Go SDK
- Configure the Vue.js SDK
- Configure the Next.js SDK
- List images in Next.js
- Lazy load images with Next.js
- Add images to Sanity with Next.js
- Image fallbacks in JavaScript
- Captioning on upload with Node.js
- Create PDFs from images in Node.js
- Delete assets with Node.js
- Manage images in a Django app
- Image gallery with Astro
- Cloudinary CLI
- Widgets and tools
- Optimization and delivery
- Transformations
- Get started with transformations
- Advanced transformation features
- Text overlay transformations
- Add watermarks in JavaScript
- Complex transformations
- Named transformations
- Advanced image components
- Trim videos in Node.js
- Splice videos in Node.js
- Zoompan effect
- Video transformations
- Crop and resize images in React
- Crop and resize videos in React
- Crop and resize images in Python
- Crop images in PHP
- Remove backgrounds and add drop shadows
- AI generative fill in Next.js
- Color accessibility in JavaScript
- Transformations for social media
- Dev Hints on YouTube
- Dev Hints en Español
- Cloudinary Café Training Sessions
- Programmatic asset management
- Additional onboarding resources
- Guides
- Cloudinary Image
- Product overview
- Image transformations
- Image transformations overview
- Named transformations
- Transformation rollout
- Image format support
- Delivery types and flags
- Transformation types
- Resizing and cropping
- Placing layers on images
- Effects and enhancements
- Background removal
- Generative AI transformations
- Face-detection based transformations
- Advanced image transformations
- Image transformation recipes
- Image optimization and delivery
- Programmatic image creation
- Product Gallery widget
- Media Editor widget
- Image add-ons
- Troubleshooting and tips
- Troubleshooting image transformation errors
- 40x errors and X-Cld-Error messages
- Images delivered without gzip compression
- Long time to first byte (TTFB) on WebPageTest.org
- Update a named transformation
- Cloudinary vs Imagga cropping
- Placeholder images (LQIP)
- How to detect corrupted images
- Force specific image when sharing links on Facebook
- Blurry images
- Cloudinary Video
- Product overview
- Cloudinary Video Player
- Video Player features
- When to use the video player
- Installation and setup
- How to embed the Video Player
- Video Player Studio
- Video Player customization
- Playlists and recommendations
- HLS and MPEG-DASH
- Video Player accessibility
- Video Player advanced features
- Video Player API reference
- Player Profiles API reference
- Video Config API reference
- Adaptive bitrate streaming
- Video transformations
- Video transformations overview
- Named transformations
- Transformation rollout
- Video Canvas
- Transcoding and formats
- Delivery types and flags
- Transformation types
- Resizing and cropping
- Trimming and concatenating
- Placing layers on videos
- Effects and enhancements
- Audio transformations
- Advanced video transformations
- Video transformation recipes
- Video and audio optimization
- Video transcription
- AI Video Analysis
- AI Video Analysis API reference
- Video analytics
- Live streaming
- Video add-ons
- Video best practices
- Social media videos
- Troubleshooting and tips
- Upload
- Asset management
- CRUD asset management APIs
- Media Library for developers
- Search for assets
- Eager and incoming transformations
- Custom metadata
- Analyze assets
- Moderate assets
- Webhook notifications
- Advanced asset management
- Add-ons
- Troubleshooting and tips
- Admin API authentication issues
- Why has the public ID of my asset changed?
- Asset ID vs public ID
- Check resource type and delivery type
- Get image or video dimensions
- Download account assets
- Delete or modify original images
- Does Cloudinary store any cookies?
- Underscores in public IDs or folder names
- Update already fetched asset
- Upload, manage, and deliver PDFs
- Cloudinary "sample" images
- Avoiding duplicate uploads
- Permitted naming conventions
- Adult/NSFW content policy
- High transformation rate
- Account management
- Retail and e-commerce
- User-generated content
- Accessible media
- AI in action
- Native mobile
- Add-ons
- Advanced Facial Attributes Detection
- Amazon Rekognition AI Moderation
- Amazon Rekognition Video Moderation
- Amazon Rekognition Auto Tagging
- Amazon Rekognition Celebrity Detection
- Aspose Document Conversion
- Cloudinary AI Background Removal
- Cloudinary AI Content Analysis
- Cloudinary AI Vision
- Cloudinary Duplicate Image Detection
- Cloudinary Image Generation
- Google AI Video Moderation
- Google AI Video Transcription
- Google Auto Tagging
- Google Automatic Video Tagging
- Google Translation
- Image to Video
- Image to Video API reference
- Imagga Auto Tagging
- Imagga Crop and Scale
- Perception Point Malware Detection
- Microsoft Azure Video Indexer
- OCR Text Detection and Extraction
- Pixelz - Remove the Background
- URL2PNG Website Screenshots
- VIESUS™ Automatic Image Enhancement
- WebPurify Image Moderation
- Cloudinary Image
- References
- SDKs
- Release Notes
How do I get a thumbnail image from a specific moment in a video?
Get Started with Cloudinary
Paste one prompt into your AI coding assistant and it sets up Cloudinary for you: installs the SDK, configures skills and MCP servers, and validates your setup.
Install Cloudinary Agent Skills
Cloudinary agent skills give your AI coding assistant first-class context for working with Cloudinary.
Run to install or update the skills:
View the Skills repo on GitHub | Learn about Cloudinary Skills
Last updated: Nov-09-2026
.jpg, and add a start offset (so_) in seconds:so_8.0/van-driving-mountain-road.jpg
The offset only affects the image. It doesn't change the video, which still plays from the beginning.
The question
"How do I change the video thumbnail to a specific part of the video, without making the video start from that point?"
A poster is the still image that shows before a video plays, and in lists and previews. By default, Cloudinary uses the middle frame of the video. Often that isn't the best frame, so you may want to choose another one.
The original asset
The examples use this 10.9-second video:
The solution
-
Request the video as an image. Use an image extension (
.jpg,.png,.webp, or.avif) in place of the video extension, to deliver a single frame as an image. -
Choose the frame. Add the start offset (
so_) as the number of seconds from the start of the video, or as a percentage of the video's length (such asso_50p).
This example takes the frame at 8 seconds:

Without an offset, the thumbnail is the frame from the middle of the video. For the first frame, use so_0. These are the frames at 2 seconds and at 50% of the video:


Let Cloudinary choose the frame
To have Cloudinary pick a suitable frame automatically, use so_auto:

Resize the poster to fit your page
Add a resize in a separate component, so that the poster is the size you need. This example creates a 700x400 poster from the frame at 8 seconds:

Use the image as the poster of an HTML video
Set the image URL as the poster attribute of your video tag, so it shows before the video plays:
so_ offset in the poster URL doesn't change the video URL. Don't add it to the video's src, or the video itself starts from that point.Related topics
✔️ Feedback sent!
Error
Unfortunately there's been an error sending your feedback.