ca3db is a "content-addressable avatar archive", intended for permanently storing multiple versions of user avatars found on social networks and websites.

ca3db exposes an HTTP API for archiving images. Once deployed, there is just one API endpoint which accepts the following parameters as a JSON payload:

POST /archive

• url: The URL of the image to store
• bucket: The name of the S3 bucket to store the image in
• region: The Amazon region that the S3 bucket lives in
• key_id: An Amazon API ID that has write access to the S3 bucket
• secret_key: The secret key for the above ID
• max_height (optional): If specified, the image will be resized to this maximum height if larger

The service will fetch the URL, and store the image in the S3 bucket. The filename is a hash of the file contents, so calling this multiple times with the same image will not store duplicate photos. Similarly, if the same URL gets replaced with a different image, both images will be stored at different URLs.

The response will be a JSON payload with the URL of the archived image, e.g.:

{
  "url":"https://s3-us-west-2.amazonaws.com/ca3db/kylewm.com/dcffbb0712bbccc3ed94fc0f0c873ce8fde83d0cc3474fff93109042c378e2f4.jpeg"
}