Get specs from anywhere. Run AI loops from zero to prod.
Terminal
Joined February 2026
- Tweets19
- Following44
- Followers106
- Likes28
ralph-starter v0.4.0 is out 🚀
What's new:
→ Visual validation — pixel-by-pixel screenshot comparison catches every CSS mismatch automatically
→ Figma upgrade — works without Auto Layout, fetches images & detects Google Fonts from your design
→ fix --design regen — re-parses your Figma spec with the latest parser so fixes use corrected instructions
→ Auto-installed skills — agents get Tailwind, shadcn, and framework skills injected automatically
Try it: npx ralph-starter@latest Release notes: github.com/multivmlabs/ralph…
Now if you build your website, application, or any tech, you can support us adding a badge to the website/repository:
check it out ralphstarter.ai/badge
ralph-starter v0.3.0 is out 🚀
What's new:
→ Smart UI defaults — Tailwind v4 + shadcn/ui auto-configured for web projects
→ fix --design mode — screenshot-driven visual QA loop that catches CSS bugs automatically
→ Smarter loop engine — better stall detection, orphan cleanup, iteration memory
Try it: npx ralph-starter@latest
release notes:
github.com/multivmlabs/ralph…
Ralph Starter retweeted
It also works with specs from:
- GitHub Issues & PRs
- Linear tickets
- Notion pages
- Figma designs
- Local files & URLs
One CLI to rule them all.
Try it: npx ralph-starter
Docs: ralphstarter.ai
Ideas repo: github.com/multivmlabs/ralph…
Ralph Starter retweeted
Imagine generating a full landing page website in less then 20 minutes, with 1 terminal command.
thats @ralphstarter - ralphstarter.ai
You run a command like:
npx ralph-starter run --from github.com/multivmlabs/ralph…
or if authenticated already on gh terminal:
npx ralph-starter run --from github --project multivmlabs/ralph-ideas --issue 86
ralph-starter v0.2.0 is here.
From beta to stable. From specs to production.
What's new:
Auto mode — batch process issues into PRs
Idea mode — AI brainstorms what to build
19 workflow presets (TDD, debugging, security audits...)
Prompt caching — up to 90% cheaper loops
Batch API — 50% cost reduction
Better CLI with progress tracking & cost display
👉 npx ralph-starter
github.com/multivmlabs/ralph…
