Yes, to a point. MistV2 works by guiding the predicted gradient in a training program towards the "mist" that is applied. When the model attempts to correct this deviation in its training, any images that are output from it will receive chaotic texturing in their sampling process, producing a terrible image.

However, as stated by the Psyker team - "Mist-v2 may only raise the cost of a motivated and professional attacker and help against someone who’s not really trying or not an expert. It is not a perfect solution but would be certainly better than nothing."

Kasumi operates on a queue system. When you submit your images, they are inserted into a queue, and are processed as a batch when it is your turn. Currently Kasumi allows you to process 25 images per job, and takes approximately 20 minutes for a job of that size to complete. You can view your job statuses from your user portal.

Yes! As long as your job has not started processing, you can cancel your job for a full refund of credits. We cannot terminate a job in processing.

Once your images are processed, you will receive an email with a Litterbox download link to a ZIP file of your images.