We are developing TAPe+ML v2, a recognition model that solves any computer vision task with the accuracy of SOTA models, but with a significantly lower number of parameters than the competitors, which have millions or even hundreds of millions of parameters.

TAPe + ML

Theory of Active Perception or TAPe are scientific language and computational methods that Comexp Research Lab has been working on for over 15 years.

Based on these methods, they have already created tools for working with video - the "heaviest" and most complex in terms of processing, and yet the fastest growing content.

Go through the tabs and explore!

Faster, more accurately and more efficiently

upload

process

result

Using TAPe techniques, we developed an "alphabet" based on the elements and structure of TAPe and built a search tree for video, TAPe-database.

Search any size video archive prepared with TAPe methods - instantly. YouTube, TikTok - you name it.

We can now work with video as conventional search engines and AI work with with text and images, while spending even less resources on video search than conventional search engines spend on text.

In the prepared and constantly updated video database you can work with different video tasks and get instant results.

a Telegram video search bot by video in the world's largest library of movies, TV shows, series and more. 

More...

the ability to use all functions of the TAPe search engine RVS for third-party professional developers and applications or without engaging professional developers.

a simple possibility any user to take advantage of TAPe API functions by customizing on their computer, laptop or smartphone the necessary usage scenarios.

If you need help please contact us or complete form