KinoPipe
Start for free
All toolsvideo thumbnail generator API

Video Thumbnail Generator API

Generate a sharp thumbnail or poster frame at an exact timestamp. Run the preset directly on this page or call the same agent-ready endpoint from your product.

Run this tool View API recipe

Prefer the shell? Extract a thumbnail from a video with FFmpeg, the verified FFmpeg command for this edit.

LIVE TOOL
Run video thumbnail generator
Drop a video, adjust the defaults if you need to, and download the result here.

or ·

Already have an account? Sign in

BEFORE / AFTER

See what the tool changes.

A real example: the after side is the unedited output of a KinoPipe job. Run the live form above to generate the same result from your own media.

BeforeSource video
00:0000:09
00:00–00:09
AfterPoster frame
Ready
00:04.0 · WebP

Demo footage: public-domain timelapses (Bureau of Land Management, Oregon · NASA SVS). Processed by the same pipeline the API and MCP tools call.

Useful defaults, typed options.

The tool slug stays stable while your agent supplies named media inputs and a narrow set of documented options.

  • Exact poster frames
  • JPEG and WebP outputs
  • Deterministic timestamps
What this preset does
  1. 01Read the source video
  2. 02Seek to the requested timestamp
  3. 03Extract one high-quality frame
  4. 04Encode a WebP or JPEG image
Stable endpointPOST /api/v1/tools/video-thumbnail-generator
Live API request
{
  "inputs": [{ "id": "main", "url": "https://example.com/video.mp4" }],
  "options": { "timestamp": 3.5, "format": "webp" }
}
Successful output example
{
  "result": {
    "output": {
      "filename": "video-thumbnail-generator-output.jpg",
      "contentType": "image/jpeg",
      "byteSize": 437021,
      "downloadUrl": "https://cdn.kinopipe.com/…"
    }
  }
}

The command this runs

ffmpeg -ss 3.5 -i input.mp4 -frames:v 1 -an -q:v 2 output.jpg
-ss 3.5 before -i
Jumps to the timestamp instead of decoding everything before it, which on a two-hour file is the difference between instant and minutes. Plenty of guides still say this costs you accuracy and makes you choose between -ss before and after the input. On FFmpeg 9.0.1, which is what runs here, the two orders produce a byte-identical image; we checked rather than repeated it.
-frames:v 1
One frame, then stop. Without it FFmpeg happily writes a numbered image for every frame it decodes.
-q:v 2
JPEG quality, on a scale where 2 is near-best and 31 is worst. The default is noticeably softer, and a poster frame is the one image a viewer judges your video by before pressing play.
-an
Drops the audio. A still has none, and some containers error rather than ignore the stream.

The hard part is choosing the frame, not extracting it

Pulling a frame is one line. Knowing which second to pull is the actual problem, and this tool does not solve it: it takes the number you give it and never looks at the picture.

If you are picking blind across a library, detect_scenes returns the timestamps where the shot changes, which are usually the frames worth showing, and find_highlights ranks moments by what is being said. Feed one of those numbers in here rather than guessing at one second and hoping.

What this will not do

It will not resize. You get the frame at the source resolution, so a 4K video yields a 4K JPEG, which is rarely what a thumbnail slot wants.

It will not add text, a play button or a border. Nothing is drawn on the frame.

It will not avoid a bad frame. Black at the start, a blurred pan, a face mid-blink: it extracts what is there at the second you named.

Doing it yourself

One line, no dependencies past FFmpeg, and worth keeping in a snippet file. The reason to call this instead is a pipeline: an agent that has just detected scenes has the timestamps in hand and can ask for the frame without shelling out, and a request for a timestamp past the end of the video comes back as a typed error rather than an empty file.

Know the boundaries before you run.

Timestamp must be between 0 and 86,400 seconds.
The timestamp must fall inside the source duration.
Available outputs are JPEG and WebP.

About generate thumbnails

You choose it, in seconds, and one second is the default. Nothing here looks at the picture, so if your video opens on black then a thumbnail at one second is black.

Yes, and you get the exact frame rather than the nearest keyframe before it. The old advice that a fast seek costs you accuracy is out of date on a current FFmpeg.

Related media tools

Convert Video to GIF Online or by API

Turn a short video section into a smooth, palette-optimized animated GIF.

Explore

Trim Video with an API

Cut a precise segment from a source video using start and end times.

Explore

Create a Social Video Clip

Cut and reframe a source video for Shorts, Reels or vertical social feeds.

Explore