Pruna AI

A model catalogue spanning text-to-video, image-to-video, reference-to-video, video-to-video, editing, replacement, upscaling and AI avatars, plus image generation, upscaling, face swap and virtual try-on.

Published figures include sub-second to a few seconds of inference for many video tasks — for example p-video-upscale at roughly 1.34s and p-video-2-pro text-to-video at about 1.99s — with text-to-video on p-video-2 in the tens of seconds.

Per-second pricing starts around $0.01 and generally lands between $0.01 and $0.05 per generated second depending on model, resolution and task.

Access hosted models through straightforward API calls, or deploy the same models within your own infrastructure via the open-source Pruna library for compliance and custom requirements.

Pruna frames optimised models as faster, cheaper, smaller and greener, reducing the compute — and therefore the energy — required per generated second.