What is uploaded and saved?

The video stays on your device. Only selected screenshots, the optional title, file name and duration are sent to our server. Screenshots are sent to Qiniu’s image model. Reference payloads are removed after provider acceptance; the first-frame preview, task and result remain for 7 days or until you delete the task. Provider retention follows its own policy.

Choose up to 4 frames, each resized to a maximum long edge of 1536 pixels and a file size of 1 MiB. Each task generates one 1536 × 864 PNG. A browser can submit 2 tasks/hour and 3/day, with one active task. Shared IP limits are 5/hour and 10/day; site limits are 20/hour, 200/day and 3 active tasks. Accepted attempts, including failures, use quota. Hourly and daily submission quotas reset at the next UTC hour and 00:00 UTC respectively.

Can I come back later?

Yes. Use the same browser and keep its cookies. Accepted tasks continue after you close the page. Tasks and images remain available for 7 days from submission; they do not sync across devices.

How are the video frames selected?

The browser extracts 24 evenly spaced frames. Choose 1–4 clear frames that represent your video, or capture the current player frame. The first selected image is the main reference; the AI does not watch the whole video.

Can I add a title or keep the original details?

A title is optional. Keep it short; leaving it blank requests no new text, while existing text may remain. AI can change faces, objects or lettering, so review the generated thumbnail before publishing.

Video to thumbnail

AI YouTube Thumbnail Creator

Local frame extractionCloud AI16:9 PNG

Choose 1–4 frames from your video, add an optional title, and generate a 16:9 YouTube thumbnail with AI.

Temporarily unavailable

This tool is taking a short break

The cloud service used by this tool is currently paused, so new files and tasks cannot be processed right now.

You do not need to upload a file. Please check back later after the service has been restored.

Your tasks

Your submitted tasks and thumbnail previews will appear here.

How it works

Three steps to finish the browser workflow.

1
Open a video
Choose a video. 24 candidate frames are extracted locally; you can also capture the current frame.
2
Choose frames and a title
Select one to four frames and optionally enter the exact title to render. The first selection is the main reference.
3
Generate and download
Generate a 1536 × 864 PNG thumbnail. Follow its progress in the task list, then preview and download the result.

Everything you need

Core details about privacy, browser support, output, and practical limits before you start.

Local frame extraction

Your video stays on your device. Only the selected reference frames are sent for AI generation.

Title (optional)

Add a short headline or leave the title blank. Check generated text before publishing.

Your tasks

Track progress, preview completed thumbnails and download them from your task list.

Input, output and limits

Check supported inputs, output formats and usage requirements before you start.

Supported input	Video files such as MP4, MOV and WebM. Container and codec support depends on the browser and processing runtime.
Output result	Export options: PNG (1536 × 864).
Processing and privacy	This service is currently paused; new processing tasks are unavailable. Frames are extracted on your device. Selected frames, the optional title, file name and duration are sent to the service; the original video is not uploaded.
Browser requirements	A browser that can decode your video locally, enough device memory, and an internet connection for AI generation.
Usage limits	Choose 1–4 reference frames per task. Each task generates one 1536 × 864 PNG; browser memory, free quotas and service capacity apply.
What is uploaded and saved?	The video stays on your device. Only selected screenshots, the optional title, file name and duration are sent to our server. Screenshots are sent to Qiniu’s image model. Reference payloads are removed after provider acceptance; the first-frame preview, task and result remain for 7 days or until you delete the task. Provider retention follows its own policy.