cclank

cclank / lanshu-awesome-ai-video-kit

Public

做企业 AI 视频项目逼出来的工具包 · 411 prompt · 15 模型 · 7 Claude Skill · 14 篇方法论

lanshu-awesome-ai-video-kit.lank.workers.dev ai-video awesome awesome-list claude claude-code
45
14
100% credibility
Found May 26, 2026 at 45 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
HTML
AI Summary

This is a comprehensive open-source toolkit for AI video prompt engineering. It contains 411 battle-tested prompts across 15 different AI video models (including Sora, Kling, Veo, Runway, and open-source options like LTX-Video and CogVideoX), organized by scene type and model. The project includes 14 methodology guides that teach the 'director language' of prompt writing - covering camera angles, lighting, emotion externalization, and model-specific formulas. Seven Claude Code Skills automate the prompt-writing process, allowing users to describe their video idea in plain language and receive a tailored prompt. A web-based browser lets users filter and search prompts without any technical setup. The project also includes a weekly auto-monitoring system that tracks official model documentation for updates. All content is sourced from official model documentation and community reviews, with proper attribution.

How It Works

1
🎬 You want to make AI-generated videos

You've heard about AI video tools like Sora, Kling, or Veo, and you want to create stunning videos. But you quickly realize that writing good prompts is an art form in itself.

2
📚 You discover a treasure chest of 411 prompts

You find a massive collection with battle-tested prompts for 15 different AI video models, organized by scene type and model. Each prompt comes with tips on what works best.

3
🔍 You browse and filter to find your perfect match

The web browser lets you filter by model, scene type, and tags. You can search for 'product commercial' or 'cinematic dialogue' and see dozens of ready-to-use examples.

4
You have two paths to choose from
✂️
Quick path: Copy and tweak

Pick a prompt, modify it for your idea, and paste it into your AI video tool. Done in minutes.

📖
Learning path: Master the craft

Read the 14 methodology guides to understand the 'director language' - camera angles, lighting, emotion externalization, and model-specific formulas.

5
🤖 You let AI help you write custom prompts

The Claude Code Skills act like a creative assistant. Tell it 'I want a product ad with a dramatic reveal' and it generates a tailored prompt using the right formula for your chosen model.

6
🔄 You compare how models handle the same scene

The cross-model matrix shows you how a 'perfume ad' prompt differs across 11 models. This helps you pick the right tool for your specific needs.

🎉 Your video comes to life exactly as imagined

Whether you copied a prompt or crafted your own, your AI-generated video matches your vision. The methodology taught you to think like a director, not just a description writer.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 45 to 45 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is lanshu-awesome-ai-video-kit?

This is a prompt engineering toolkit for AI video generation. It bundles 411 battle-tested prompts across 15 different models—including Sora, Veo, Kling, Runway, and several open-source options like LTX-Video and CogVideoX—along with 14 methodology guides explaining how to write effective prompts for each platform. The project also includes seven Claude Code Skills that automate prompt generation, letting you say "make a product ad video" and have Claude pick the right model and write the prompt for you. A web-based browser lets you filter and search prompts by model, scenario, and tags without touching any markdown files.

Why is it gaining traction?

The AI video space moves fast, and prompts that worked six months ago may fail today. This project solves that with automated monitoring—every Monday, a GitHub Action checks 32 official endpoints and alerts maintainers when models update. The cross-model comparison matrix is particularly useful: it shows how the same "perfume ad" prompt renders differently across 11 platforms, so you can pick the right tool for your budget and use case. The Claude Code Skills integration removes the learning curve—you don't need to memorize 15 different prompt formulas when Claude can handle it.

Who should use this?

Content teams producing marketing videos who need to pick between Kling, Sora, or Runway without running expensive A/B tests. Developers building AI video pipelines who want structured prompts rather than trial-and-error. Anyone working with Chinese-language video generation—the project has strong coverage of models like Kling, Jimeng, and Wan that Western tools often overlook.

Verdict

At 45 stars, this is a niche but well-maintained resource. The 1.0% credibility score reflects its small community, but the sourcing is solid—prompts cite official vendor documentation, not random blog posts. If you're serious about AI video production, the methodology docs alone justify a bookmark. Start with the prompt browser to find templates, then install the Claude Skills to let the tool do the heavy lifting.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.