Blurhash

BlurHash applies a simple DCT transformation to the image data, retaining only the first few components, and then encodes these components using base 83 encoding with the JSON character set, HTML:

eTGu,l%Mxu-;t7~qxut7t7ofRjWBt7IUWBRjofWBRjj]-;WCIoxut7

You can then perform a reverse conversion to sRGB space over the string.

You can read more about the algorithm here.

Example of image conversion

Service life cycle

You make an http-request to the service, passing in the url of the request a link to your image, as well as x and y parameters (values from 1 to 9) - the number of displayed components.

How does it work?

Request

http://localhost:3333?image_url=https://imgproxy.com/imageurl&x=1&y=4

Response

{
  "blur_hash": "eTGu,l%Mxu-;t7~qxut7t7ofRjWBt7IUWBRjofWBRjj]-;WCIoxut7"
}

How to deploy the service

docker

docker run -d -p 3333:3333 ghcr.io/dev-family/blurhash:1.0.0

docker-compose

blurhash:
  image: ghcr.io/dev-family/blurhash:1.0.0
  ports:
    - "3333:3333"