Grok Imagine in Action: A Textbook Structured Prompt for Text-to-Video
tetsuoai · x · 2026-08-22
A user shared the full text-to-video prompt written for Grok Imagine: a 15-second locked-off shot in a cluttered gaming apartment where an anime woman knocks out a gamer with a frying pan and runs off giggling. The value lies in the structure: the prompt is broken into scene context, a location map (foreground/midground/background), first frame & blocking, format mode (one continuous shot, no cuts), optics (47° wide FOV, deep focus, no drift), and camera (locked tripod), followed by a timeline describing each action beat to 0.5-second precision (sneak-up → wind-up → 0.4s swing at 30 km/h → knockout). A solid reference template for text-to-video prompt engineering.
More from Multimodal
- fal launches Worlds for real-time creation and interaction with AI-generated worlds — Scobleizer · 2026-10-16
- Gummy worms as a physics test: Gemini 4 Argon vs. Claude Opus 5.5 elasticity sims — DynamicWebPaige · 2026-10-16
- 260M-parameter looped diffusion transformer beats 6x larger baselines with 5x less compute — burkov · 2026-10-16
- Seedance 2.5 demo: stretching one second of an explosion into a 30-second shot — LudovicCreator · 2026-10-16
- Octopus Short Film Made with Grok Imagine, Upscaled via Topaz — creatoroff · 2026-10-16
- Open-sourced 61-second computing history film generated entirely by Muse Spark — sophiamyang · 2026-10-16