- everything is local until you deploy
- with first-class support for
- connect your AI to playground
- version-controlled deployments
- full playground control
- output viewport size, zoom, color scheme control
- output eval for AI for immediate tests and feedback
- output live capture and image-cropped element/region pins
- output live stream picture-in-picture to watch your agents work
- state preserving 2-way IO sync
- multi-level open shadow root IO sync
- service/edge workers in the playground (develop full-stack apps in one place)
- import ESMs or WASMs by URL (yes in workers!)
- examples: https://new.rt.ht
- example: https://post.rt.ht
- example: https://kv.rt.ht/fs
- enterprise-class global network: https://RTEdge.net - built-in auth for deployments: https://elefunc.com/#auth
Too many features to list them all here! A lot of these features are unique and will never exist in other places, because everyone is busy chasing the next release of React or other cancerous bullshit instead. I only use web standard/platform APIs and drive things up to their simplest forms! Auth is just 4 letters added to location.pathname for example!These days, I don't even code in https://RTCode.io myself, I share the playground with Codex/Claude and let them build me my dreams in the form of full-stack apps! I PiP-watch them live on the side as they work and deploy!
For more info: https://elefunc.com/#code
Probably harder, and wouldn't work equally well for everyone, but it would be possible to have "guess language and autocomplete" entirely client-side, especially when it comes to JavaScript of all languages.