• Docs >
  • Discover what Lightning Apps can do in 5 min
Shortcuts

Discover what Lightning Apps can do in 5 min

Lightning Apps can be plenty things, and while a picture is worth a thousand words, videos showing you examples should be worth even more.


Flashy - Auto ML App (Public)

Train a model on any image or text dataset without writing any code. Flashy uses React.js for its frontend.

Find Flashy on the App Gallery and the Flashy codebase. on GitHub.


Research App (Public)

Share your paper bundled with the arxiv link, poster, live jupyter notebook, interactive demo to try the model, and more!

Find the Research App on the App Gallery and the Research App codebase. on GitHub.


ScratchPad - Notebook Manager for Team (Public)

Run multiple Jupyter Notebooks on cloud CPUs or machines with multiple GPUs.

Find the ScratchPad App on the App Gallery and the ScratchPad App codebase on GitHub.

Note

ScratchPad is tested end-to-end on every Lightning App commit with pytest.


InVideo Search (Public)

This App lets you find anything you’re looking for inside a video. The engine is powered by Open AI CLIP.

Find the InVideo Search App on the App Gallery and the InVideo Search App codebase. in GitHub.


AI-powered HackerNews (Public)

Save yourself time, and get Hacker News story recommendations, chosen for you specifically. This Lightning App was designed to illustrate a full end-to-end MLOPs workflow aimed at enterprise recommendation systems.

Find the AI-powered HackerNews App on the App Gallery and the AI-powered HackerNews App codebase. on GitHub.


Lightning Apps can turn ML into scalable systems in days — not months

Use the Lightning framework to develop any ML system: train and deploy a model, create an ETL pipeline, or spin up a research demo — using the intuitive principles we pioneered with PyTorch Lightning.

Apps with Logos

Anyone who knows Python can build a Lightning App, even without machine learning experience.

Lightning Apps are:

  • cloud agnostic

  • fault-tolerant, distributed, cost optimized

  • production ready

  • local and cloud debuggable

  • highly reactive & interactive

  • connect multiple UIs together

  • built for team collaboration

  • framework agnostic, use your own stack

  • and much more




Next Steps