AI video streets turned into walkable 3D spaces with GPT6 Astra
Dr_Singularity · x · 2026-09-05
A Japanese developer demoed a workflow chaining Midjourney + Kling-generated street videos into GPT6 Astra, which produced a walkable 3D space based on the AI video.
It's a first-pass result with rough textures, but the author argues this video-to-3D pipeline opens a direct path toward worldbuilding and IP creation. The how-to is shared in the replies.
More from Multimodal
- Seedance 2.5 'ugly filter to glam' video with full shot-by-shot prompt released — SimplyAnnisa · 2026-09-05
- Gave up local video rendering on an RTX 3060 for cloud — but lost the node-based modularity — Suspicious_Pizza9529 · 2026-09-05
- GPT-6 Astra drawing Hatsune Miku turns heads on X — FateOfMuffins · 2026-09-05
- Runway ships Solaris interface world model; PSI exits stealth with $58M seed — shashib · 2026-09-05
- Runway CEO unveils Solaris, a real-time interactive "interface world model" video model — c_valenzuelab · 2026-09-05
- ComfyUI hogs all VRAM: --reserve-vram broken, undocumented --vram-headroom no help — Radiant-Photograph46 · 2026-09-05