ziplab / ATAS
git clone https://github.com/ziplab/ATAS.git
Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generation
+1
stars, 24h · #9,975
+~7
stars, 7 days · #10,623
+303
stars, 30 days · #1,688
323
★ total · +2% this week
5
forks · 0 open issues
Star history (as recorded here)
LanguagePython
LicenseMIT
Created2026-10-12 (10d ago)
Last push7d ago
Watchers3
OwnerOrganization
GitHub topics
agentic-video-generationmutli-agentmutli-agent-systemspatial-consistencyvideo-generationvideo-generator
Links
- GitHub: github.com/ziplab/ATAS
- Homepage: https://lhmd.top/atas
- Releases · Issues · Hacker News mentions
README excerpt
Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generation Weijie Wang 1,* Haomin Zhang 1,* Ye Li 1,* Akide Liu 2 Zeyu Zhang 3 Haoyu Zhao 1 Yefei He 1 Donny Y. Chen 2 Feng Chen 4 Bohan Zhuang 1,† 1 Zhejiang University 2 Monash University 3 BAIR, UC Berkeley 4 Adelaide University * Equal contribution. † Corresponding author. **ATAS keeps one persistent world behind a whole film.** Every shot is generated by projecting the world state into the requested view, and every generated shot is reconstructed, registered and validated before it updates that world. ## Highlights 🧭 World-State Compiler A screenplay becomes bounded shot requests with explicit scene, camera, time and event. 🏠 Scene Memory Static geometry of every place in one shared frame, rendered for any revisit from any viewpoint. ⏱️ Event Memory Dynamic observations indexed by scene, event and time, so a past moment can be re-filmed from a new angle. ✅ Validated Updates The world changes only when film review, registration and the state check all pass. ## Method A world-state compiler turns the screenplay into shot requests; each request queries the scene and event memory, the rendered context conditions the video generator, and the new clip is reconstructed and registered before it can enter the world. ## Results Human evaluation against multi-shot baselines (1–5). Spatial revisit and event recapture. Runtime, candidate lifecycle and delivery efficiency across memory configurations. ## Installation The client needs no GPU; the geometry service needs an NVIDIA GPU (about 20 GB free memory). [code] `GEOMETRY_MODEL_ROOT` contains `pi3x/` (Pi3X) and `sam3/sam3.pt` (SAM3). ## Usage [code] Any plain-text screenplay works (`--characters` is optional). Other options: `--memory none` (no world state), `--plan-only`, `--max-shots N`, `--output DIR`. Re-running the same command resumes without repeating paid generations. Results go to `client/runs/ /story_ /`: `final.mp4`, a per-shot `manifest.json` and `report.md`. Examples in `examples/`: `bookstore`, `clockshop`, `record`. Design notes: docs/design.md. Tests: `cd client && python -m pytest`, `cd geometry && python -m pytest`. ## Citation [code] ## Acknowledgements ATAS builds on Pi3 and SAM 3. We thank the authors for releasing their work. ## License This project is released under the MIT License. Third-party code under `geometry/backbones/` keeps its own license files.