About
What is trace.moe?

trace.moe is an Anime Scene Search Engine that helps users to trace back the original anime by a screenshot. It search in ~30000 hours of anime and find the best matching scene. It can tell the anime, the episode and the exact time that scene appears. Since the search result may not be correct, it provides a few seconds of preview for verification. There has been a lot of anime screencaps and GIFs spreading around the internet without quoting the source. And trace.moe is built to fix that, helping people to get to know the source anime, not just some random piece of work in content farms.

trace.moe is a free service and has no Ads. It relies entirely on donations for its operational costs.

What trace.moe is NOT

This website is not for watching anime. The server has effective measures to prevent users from accessing the original video beyond the preview limit. If you ask me where you can watch the anime in search result, I'll show you the way to Youtube, Amazon, Netflix, bilibili, etc (depending on your country).

trace.moe is not for comics / anime-style artworks. This search engine only index anime officially published through TV/Web/DVD/Bluray. If you wish to search artwork / wallpapers, try to use SauceNAO and iqdb.org

trace.moe is not an AI. It does not have a neural network that recognize and understand the things (like characters) on the images. It uses a technology called Content-based image retrieval which compares only the colors and patterns of the images instead of trying to understanding the image. Thus, it is nothing related to Machine Learning and is not train-able. You may read the presentations slides for technical details.

System Status

This search engine index latest anime automatically, usually within hours after broadcast. RSS Feeds

  • Analyzed Video: counting...
  • Total Duration: counting...
  • Analyzed Frames: counting...
  • Indexed Frames: counting...
  • Index Size: calculating...

Last Database Update:

Check database coverage by Anilist ID:

Credit
Dr. Mathias Lux (LIRE Project)
Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval – An Extensible Java CBIR Library. In proceedings of the 16th ACM International Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008 Visual Information Retrieval with Java and LIRE
Josh (Anilist) and Anilist team