cclank/lanshu-awesome-ai-video-kit

📚 Quick Overview

lanshu‑awesome‑ai‑video‑kit is an open‑source “prompt‑engineering” toolbox for AI‑generated video. It collects, curates and documents hundreds of high‑quality prompts for dozens of commercial and open‑source text‑to‑video models, adds conversion utilities (Claude Skills) and a few lightweight web tools for browsing the data. The repo also runs a weekly GitHub Action that checks the official model endpoints for changes and automatically opens issues when a prompt may need updating.

---

🎯 What the project actually does

Feature	How it works (as described in the README)
Prompt collection	594 vetted prompts, each with a source field linking to the original documentation or blog. Prompts are grouped by model (22 models, 15 commercial + 4 open‑source) and stored in prompts/ with a master JSON index (all‑prompts.json).
Cross‑model matrix	110 “matrix” entries that map the same use‑case across different models, stored in prompts/data/cross‑model‑matrix.json.
Claude Code Skills	Seven ready‑to‑install Claude Skills (e.g., model-selector, prompt‑translator, model‑specific promper tools). Each skill lives in its own folder with a SKILL.md description and can be linked into ~/.claude/skills/ with a single ln -s command.
Methodology SOPs	21 markdown documents that explain prompting formulas, best‑practice checklists, and “masterclass” guides for specific models (Seedance, Kling, HappyHorse, etc.).
Web UI	Zero‑dependency, single‑file HTML pages (tools/prompt-browser, tools/cross‑model, viewer.html) that run on GitHub Pages or any static host. Features include theme switching, keyboard navigation and shareable URLs.
Local dev server	serve.py starts a tiny HTTP server that adds UTF‑8 handling, automatic markdown redirects and disables caching for faster iteration.
Automatic monitoring	A GitHub Action (model-version-monitor.yml) runs weekly, calls scripts/monitor_models.py to fetch 32 official model endpoints, compares SHA‑256 hashes, and opens an issue when any endpoint changes.
Contribution workflow	Issue templates for adding prompts, uploading test videos, reporting new model versions, etc., plus a CONTRIBUTING.md guide.

---

🛠️ Who might use it?

• Product teams building AI‑video features who need reliable, up‑to‑date prompt templates for multiple providers.
• Prompt engineers looking for a structured reference (formulas, constraints, example prompts) across commercial and open‑source models.
• Developers who want quick access to a prompt‑to‑model conversion utility via Claude Skills.
• Researchers studying the evolution of text‑to‑video prompting conventions.

---

📦 Getting started (5‑second guide)

git clone https://github.com/cclank/lanshu-awesome-ai-video-kit
cd lanshu-awesome-ai-video-kit
python3 serve.py 8000   # starts a local server
# Open http://localhost:8000/ in a browser to explore the prompt browser

---

📂 Repository layout (high‑level)

├─ README*                # project description (Chinese & English)
├─ serve.py                # tiny dev server
├─ prompts/                # all prompt data + per‑model READMEs
│   ├─ data/all‑prompts.json
│   └─ data/cross‑model‑matrix.json
├─ methodology/            # 21 SOP markdown files
├─ skills/                 # Claude Skills (model‑selector, translator, etc.)
├─ tools/                  # single‑file HTML browsers
│   ├─ prompt‑browser/
│   └─ cross‑model/
├─ scripts/                # monitoring script & endpoint list
└─ .github/workflows/      # weekly monitoring action

---

🔗 Useful links

• Live demo – https://lanshu-awesome-ai-video-kit.lank.workers.dev
• English README – README.en.md
• Contribution guide – CONTRIBUTING.md
• License – MIT (prompt content is for research/education only)

---

All details above are taken directly from the repository’s README; no additional features have been inferred.