Is my video uploaded when I use the reader on this page?

No. The reader is JavaScript that runs in your browser. It plays the video from your device in a hidden video element, draws frames to a canvas and makes the JPEGs there. Nothing is sent to Mixpeek or anyone else, and once the page has loaded it keeps working with the network switched off.

How many keyframes will I get?

One for the first sample and one for every scene change the reader detects, up to 48. A talking-head recording with no cuts gives one or two. A fast-cut trailer gives dozens and can reach the limit of 48, which the result says when it happens.

What counts as a scene change?

A sample whose 32 by 18 pixel thumbnail differs from the sample before it by at least 20% in average color. Hard cuts clear that easily, including a cut between two shots of the same brightness. A slow fade or camera pan changes the picture a little at a time and can pass without a keyframe, and a shot shorter than the gap between samples can be missed. The result shows the gap it used.

Which video formats work?

Whatever your browser can play. An MP4 with H.264 video plays in Chrome, Edge, Firefox and Safari, and WebM plays in Chrome, Edge and Firefox. A MOV file plays when it holds H.264. When the browser cannot decode a file, the reader says so.

What image format are keyframes saved in?

JPEG at 90% quality, scaled down to 640 pixels wide when the video is wider. The JSON download lists each keyframe's timestamp, how much the picture changed and its average color.

How do I extract keyframes from a whole video library?

The API code below creates a bucket and a collection that runs an extractor over every video in it on Mixpeek's servers, reading the files from your storage and storing the output as documents you can search. That is the path for thousands of hours of video.

media

Video
Images
Converter

Find the scene changes in a video and save a frame from each, right on this page. The video plays from your device in your browser, nothing is uploaded, and every keyframe downloads as a JPEG with its timestamp. The page also covers doing this across a video library through an API.

Max file size: No set limit; the video plays from your device

Estimated: Up to about 30 seconds, in your browser

3 input formats

Find a video's keyframes here

Runs in your browser. The file stays on your device and nothing is uploaded.

Drop an MP4, WebM or MOV file here, or choose one from your device.

Choose a file

Loading the reader.

How It Works

Drop an MP4, WebM or MOV file onto the reader at the top of this page, or choose one. The video plays from your device in your browser.

The reader samples up to 240 frames spread evenly across the video, at least a quarter of a second apart.

Each sample is shrunk to a 32 by 18 pixel thumbnail and compared with the sample before it. When the average color difference reaches 20%, that sample becomes a keyframe.

Each keyframe is drawn at up to 640 pixels wide and saved as a JPEG you can download, labeled with its timestamp and how much the picture changed.

For a whole library, the API code below runs an extractor over every video in a bucket on Mixpeek's servers and stores the output as documents you can search.

Code Examples

import os, requests

API = "https://api.mixpeek.com"
H = {"Authorization": f"Bearer {os.environ['MIXPEEK_API_KEY']}",
     "X-Namespace": os.environ["NAMESPACE_ID"]}

# 1. a bucket, with a schema that declares the field you will send
bucket = requests.post(f"{API}/v1/buckets", headers=H, json={
    "bucket_name": "video-inputs",
    "bucket_schema": {"properties": {"video": {"type": "video"}}},
}).json()

# 2. land the file as an object. the URL goes in data, on the blob
requests.post(f"{API}/v1/buckets/{bucket['bucket_id']}/objects", headers=H, json={
    "key_prefix": "run-1",
    "blobs": [{"property": "video", "type": "video",
               "data": "https://example.com/clip.mp4"}],
})

# 3. a collection over that bucket, running the extractor
collection = requests.post(f"{API}/v1/collections", headers=H, json={
    "collection_name": "video-to-keyframes",
    "source": {"type": "bucket", "bucket_ids": [bucket["bucket_id"]]},
    "feature_extractor": {"feature_extractor_name": "multimodal_extractor", "version": "v1"},
}).json()

# 4. run extraction over the bucket
requests.post(f"{API}/v1/buckets/{bucket['bucket_id']}/batches", headers=H, json={
    "collection_ids": [collection["collection_id"]],
    "auto_submit": True,
})

# 5. read the output
docs = requests.get(
    f"{API}/v1/collections/{collection['collection_id']}/documents", headers=H
).json()
print(docs)

Use Cases

Build visual indexes for video libraries

Generate storyboards for film and advertising review

Create thumbnail galleries for e-learning platforms

Power visual search across video catalogs

Supported Input Formats

MP4

WebM

MOV

Quick Info

Categorymedia

Max File SizeNo set limit; the video plays from your device

Est. TimeUp to about 30 seconds, in your browser

Extractorvideo-descriptor

Processing millions of hours of video?

Run this as a managed pipeline over your entire video archive, no infrastructure to build or maintain. Talk to us about processing at scale.

Run it over a library

Mixpeek runs this conversion as a pipeline over a whole library in your object storage, with the output landing as queryable documents. The reader on this page handles one file at a time.

Frequently Asked Questions

Related Converters

Video

Thumbnails

Video to Thumbnails

Pull 12 evenly spaced thumbnails from a video right on this page. The video plays from your device in your browser, nothing is uploaded, and each thumbnail downloads as a JPEG. The page also covers generating thumbnails across a video library through an API.