AI video prompts, with the videos they made

Browse 1945+ community prompts for Seedance 2, Veo 3.1, Sora 2, Kling 3.0, Grok, Wan, and Hailuo. Open one to see the model and settings behind the video, then recreate it in Video Studio. For a focused starting point, browse the Veo 3 prompt examples.

320 results · Clear filters

Pick a model

Lowest price per run. Length, quality and audio change the final cost. Credit packages

Seedance 2.0 Text, first/last-frame, and multimodal reference video 76 cr ~EUR 0.69 Seedance 2.0 Fast Faster Seedance 2 workflow with the same multimodal options 62 cr ~EUR 0.56 Kling 3.0 3-15s video, audio and multi-shot 60 cr ~EUR 0.55 Veo 3.1 Lite Newest Veo 3.1 low-cost mode for text/image-to-video 40 cr ~EUR 0.36 Veo 3.1 Fast Google Veo 3.1 fast model 100 cr ~EUR 0.91 Sora 2 Sora 2 (10s/15s) 80 cr ~EUR 0.73 Grok Imagine Video xAI Grok text/image to video 10 cr ~EUR 0.09 Wan 2.7 Video Text-to-video generation (2-15s, 720p/1080p) 32 cr ~EUR 0.29 Wan 2.7 Image to Video Image-to-video with first/last frame control (2-15s, 720p/1080p) 32 cr ~EUR 0.29 Wan 2.7 R2V Reference-to-video with character/voice consistency (2-10s, 720p/1080p) 32 cr ~EUR 0.29 Hailuo 2.3 Standard Image-to-video (6s/10s, 768P/1080P) 30 cr ~EUR 0.27 Kling 2.6 Image/text to video with optional audio 55 cr ~EUR 0.50

What is the UlazAI AI Video Prompt Directory?

This directory is a searchable library of community-tested AI video prompts for the models currently available in UlazAI. It now reflects newer routes like Seedance 2, Kling 3.0, Wan 2.7, Hailuo 2.3, plus established workflows for Veo 3.1, Sora 2, and Grok Imagine.

How to Use These AI Video Prompts

  • Search by subject, style, model name, or prompt tag
  • Open the prompt detail page and copy the prompt text
  • Choose text-to-video, image-to-video, reference-video, or audio-capable generation
  • Check credit estimates before rendering and iterate inside the matching studio

Model choice matters. Use Seedance 2 for multimodal references, Kling 3.0 for motion-heavy clips, Veo 3.1 for polished prompt fidelity, and Wan/Hailuo routes when your input type or budget fits better.

Popular Categories

Related Prompt Resources

External references: Google DeepMind Veo, OpenAI Sora overview, OpenAI API docs.

Frequently Asked Questions About the UlazAI Prompt Directory

How the directory, prompt assistant, model routes, and credit estimates work together.

What is the UlazAI Prompt Directory?

It is a searchable collection of 1945+ free AI video prompts for Seedance 2, Veo 3.1, Sora 2, Kling 3.0, Grok, Wan, and Hailuo. Copy a prompt, adapt it, and open the matching model in Video Studio.

How can I create better prompts?

Use the AI Prompt Assistant to expand rough ideas into structured prompt drafts. Then adapt the prompt for the selected model: reference-heavy prompts fit Seedance 2 or Wan R2V, motion-heavy prompts can fit Kling 3.0, and polished cinematic prompts often fit Veo 3.1.

Which model should I choose?

Start with your input type. Text-only ideas can go to Veo 3.1, Sora 2, Kling 3.0, Seedance 2, or Wan 2.7 T2V. Image-to-video prompts fit Seedance 2, Veo 3.1, Wan 2.7 I2V, Hailuo, Grok, and Kling. Audio or multimodal reference workflows are strongest candidates for Seedance 2 or Kling routes.

How do I use a prompt from the directory?

  1. Search by model, theme, or category.
  2. Open a prompt detail page and copy the prompt.
  3. Pick the right UlazAI model route for the input type.
  4. Review credits, duration, aspect ratio, and quality settings.
  5. Render, compare output, and save the best version.

Are the prompts free?

Yes. The directory prompts are free to browse, copy, and adapt. Rendering videos uses UlazAI credits; each model lists a starting credit estimate before you open a studio.

Can I submit my own prompts?

Yes. Submit high-quality prompts with clear output intent and, where possible, an example video. Good submissions improve the directory and make it easier for other creators to understand which prompt patterns work across current models. Share a prompt.

Find a prompt by model, scene and camera move

Search the directory for the result you want, then open a prompt that matches the available model and input type. Copying a cinematic line into the wrong studio can fail even when the wording looks strong.

Read the full prompt before using it. Replace subject and brand details, but keep the shot structure, timing and camera instructions together.