Skip to content

Build a pipeline of AI models

Image, video, audio and text are nodes of one graph. Change one node — only it and what follows is recomputed.

You can build a graph without signing up. Sign-in is needed to run it.

Pay for what changed

Drops are charged for every node that runs. Finished results stay in the cache, and a good result can be pinned so it is never recomputed.

Example: the “Product card video” workflow. The numbers are illustrative — they show the calculation, not prices.
NodeFirst runChanged the background descriptionChanged the voice-over text
Remove background2cachedcached
New background88cached
Upscale3pinnedpinned
Animate1818cached
Voice-over4cached4
Add sound111
Total, drops36275

The same drops as in the bot. One account, one balance.