Research framework for long, consistent video generation using a persistent 3D world state as memory.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Any Time, Any Space (ATAS) is a research framework for agentic video generation, described in the paper "Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generation" (Wang et al., 2026). It keeps an explicit world state, reconstructed as a 3D environment, that serves as memory across a long video.\n\nAccording to a YouTube creator summary, this memory keeps long generated videos consistent, so that rooms look the same when a scene is revisited. The same summary says it works with different underlying video models and that code has been released. The project page shows video results, a world-state timeline, an example scene ("Saturday Laundry in the Flat"), and comparisons with other frameworks.\n\nATAS is a research project rather than a hosted product. The provided material does not give a repository link, license, or details on model weights, so those are not stated here.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
System that keeps long generated videos consistent by reconstructing a 3D environment as memory, so rooms stay the same when revisited. Works with different video models; code released.
Closest neighbours by meaning, not just by category.
Model
Research project for real-time autoregressive generation of multi-shot video narratives from shot-level prompts.
Model
Open framework and dataset for turning video diffusion models into real-time, long-horizon interactive world models.
Model
Real-time interactive world model with compressed memory and hybrid action and semantic control, built on Wan 2.2.
Model
Subject-driven text-to-video model that keeps multiple reference characters and objects consistent across art styles.
Model
Interactive autoregressive world model generating minute-long playable video with real-time camera control and prompt events.
Research · Meta
Meta/Oxford multi-agent video world model that adds fixed auxiliary cameras to keep a shared world consistent.
