Fluffle is an initiative taken with the goal to create platform for fellow furries with which they can reverse search artwork on a variety of sources. It is an open source project licensed under the MIT license. If you'd like to check it out, the project can be found on GitHub.
When you reverse search an image on Fluffle, the image gets stored temporarily on our servers for processing. Processing only contains fingerprinting the image. That is creating an identifier which can be used to compare indexed images. Once the fingerprint is created, the submitted image gets permanently deleted immediately. The aforementioned fingerprint also doesn’t get stored and neither do the settings used for reverse searching. There is an exception to this rule and that is when you create a permanent link to the search result. At that point, the used image is stored indefinitely on our servers and so are the results of your search query. After all, these are needed when someone opens your generated link.
The settings used when reverse searching are stored in your browser its so-called local storage. You can think of this as a place in which websites can store a small amount of data. These preferences are only sent to Fluffle whenever you submit an image for reverse searching and are immediately discarded when they are not required anymore for fulfilling your request.
However, a few things are logged regarding your request. These are strictly used for assessing server performance and request throttling/blocking. The following information is being logged:
- The origin of the request, needed to identify malicious clients and block them.
- User Agent, for a normal user, this would be the web browser and the version of said web browser. Also used to identify malicious clients. Mainly used to identify which applications are making use of Fluffle its API.
- If your request caused the server to error out for whatever reason, the corresponding stack trace of the occurred exception gets logged. This information is used to determine what caused your request to fail if you decide to report the issue.
- The format in which the image is encoded and its width and height. These three image attributes correlate highly with the time needed to process your request. It is therefore used to assess server performance. Its also used to identify API clients that don’t follow the guidelines put up by Fluffle.
- The moment at which the server started processing your request. This can be used to assess server performance in scenarios where there are a lot of request being processed in parallel.
- The how manyth request it is since the start of the server. Used to assess server performance.
- A bunch of timings. You can compare this to having a stopwatch and tracking laps, but in application code. How long did it take to create the fingerprint? How long did it take to check if the submitted file was an image? How long did it take to compare the fingerprint? Etc. Used to assess server performance.
- The number of images compared against. Used to assess server performance.
- Per category (exact, toss-up, alternative, unlikely match) the number of images in the search result. Used to assess how effective Fluffle is.
Tracking & cookies