SamuelQZQ

一个长达 10 小时的 AI 全自动编程实验

481
164
89% credibility
Found Feb 18, 2026 at 265 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A web platform that transforms user-submitted stories into animated videos by AI-generating scene descriptions, images, and short clips.

How It Works

1
🌐 Discover the story-to-video maker

You land on a simple homepage that promises to turn your stories into beautiful animated videos with AI magic.

2
👤 Sign up or log in

Create a free account with your email or log back in to start creating your own video projects.

3
Start a new project

Give your story a title, type in your tale, pick a fun style like cartoon or realistic, and hit create.

4
🎬 Watch AI break it into scenes

AI reads your story and instantly creates a sequence of vivid scene descriptions, ready for you to review.

5
Tweak and approve scenes

Read through each scene, make any changes if needed, and confirm you're happy with the breakdown.

6
🖼️ Generate stunning images and videos

With one click, AI creates beautiful images for each scene, then turns them into smooth short video clips.

🎉 Download your complete video story

Enjoy your fully animated video project, download all clips, and share your story brought to life!

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 265 to 481 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 auto-coding-agent-demo?

This TypeScript demo showcases a fully autonomous AI agent building an entire Next.js web app from scratch over 10 hours. Users get setup scripts that loop an agent through tasks in a JSON file, resulting in a complete story-to-video platform: input a story, auto-generate scenes via AI, create styled images, then short video clips—all with Supabase auth, storage, and API endpoints for batch generation. It solves the "can agents really code production apps?" question by delivering a runnable app with user flows for projects, previews, and downloads.

Why is it gaining traction?

Unlike toy agent demos, this auto-coding agent tackles real complexity: multi-stage pipelines (scenes → images → videos), error handling, polling for async AI jobs, and UI with confirmations per stage. Developers dig the bash automation script that runs agent sessions in loops, tracking progress via task.json and commits—proving scalable agentic workflows without constant human intervention. The hook? Fork it to experiment with your own auto-coding agent on TypeScript stacks.

Who should use this?

AI agent builders testing Claude integrations for app generation. Full-stack devs prototyping AI media tools like story-to-video pipelines. Researchers benchmarking autonomous coding on realistic tasks with auth, DB, and external APIs (Volc/Zhipu).

Verdict

Intriguing proof-of-concept for agent-driven development—clone and run the scripts to see auto-coding in action, especially if you're into TypeScript AI experiments. With 245 stars and 0.9% credibility score, it's raw (light docs, no tests), so treat as inspiration rather than deploy-ready; add your own polish for serious use.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.