awesome-video-prompts
What they did
A prompt collection for Veo 3/3.1, Kling and Hailuo, organised by scene type.
Copy prompt
Cinematic long shot with forward tracking: In a frozen wasteland under a black sky filled with auroras, a small squad treks across an ice ridge, their footsteps crunching over ancient wreckage. Wind howls. One pauses, raising a scope. In the valley below: dozens of beasts, dormant, coiled around a shattered mech carrier. The camera slowly tracks forward as the squad descends each step heavy, uncertain. Then one of the creatures stirs. Its eyes glow. Others follow. The ice begins to crack beneath their feet. The camera pulls upward as all hell breaks loose, beasts charging up the slope, soldiers scrambling, rifles lighting up the darkness in staccato bursts.Excerpt credited to songguoxs, from the source README.
Worth knowing
- Price: Open source repo (license unknown)
- Status: sourced. We have not run this ourselves yet (no "Tested" badge).
- Quality bar (computed by code): 5/5 (primary source ✓, reproducible ✓, new ✓, example output ✓, safe ✓)
The best new AI video workflows and prompts, weekly.
Related
▶ demo at sourceKling AI
Kuaishou's own Kling studio, where you make text-, image- and reference-driven video and images in one web app. It's the place to try new Kling versions first.
▶ YouTube demoVeo (Google DeepMind)
Google DeepMind's page for the Veo video model family, with capability notes and sample clips straight from the lab.
▶ demo at sourceVeo 3 on fal
Veo 3 served through fal's API, with text-to-video and optional generated audio. Price depends on whether audio is on.
▶ demo at sourceVeo 3 Fast on fal
A faster, cheaper Veo 3 variant on fal. It's a sensible default for drafts before you spend on the full model.
▶ YouTube demoVeo on Gemini API docs
Google's developer docs for generating video through the Gemini API, including when to pick Veo 3.1 versus faster editing-oriented models.
Google Flow
Google's filmmaking app built on Veo and its image models. You assemble shots and scenes in one place instead of generating clips one by one.
