Phenaki

By Google

Phenaki treats video as a narrative conditioned on changing text, so you can describe a scene, transition the prompt, and watch the visuals evolve in sync with your script. A tokenized video representation lets the system plan motion and scene changes across many seconds while preserving identity and style. It’s suited to multi-beat concepts—establishing shots, action, and cutaways—with smoother temporal continuity than early short-clip models. The result is a tool for drafting storyboards and explainer sequences where the prompt itself acts as a timeline.

New Video Gen 1

Released: October 1, 2022

Overview

Phenaki is a text-to-video model designed for long, story-like generations. It follows a sequence of prompts over time, stitching scenes into a coherent, evolving clip.