Anything that reads perma.sh without running JavaScript (many AI crawlers, link unfurlers in chat apps, some search snippets) gets an empty page whose only sentence is "These legal documents require JavaScript. Email [email protected] for a copy." That noscript text was written for the legal pages but is served on the homepage and every path. There is also no robots.txt, no sitemap, no og:image, and every made-up URL returns 200. For a brand-new product whose growth will come from search, AI answers and shared links, the first thing machines learn about perma.sh is wrong.
First action
View the page source of perma.sh and change the noscript line today.
How to measure it
Homepage indexed with correct snippet
- Now
- –
- Target
- Yes, in Google and Bing
- By
- 30 days
Steps
Fix the noscript
Homepage text that says what perma.sh is, with a link to the app.
Founding Engineer 1 hour
Pre-render
Static HTML for / in English and a Spanish version.
Founding Engineer 2 days
Add robots and sitemap
List every real page.
Founding Engineer 1 hour
Share image
An og:image with the price promise.
Growth Marketer 2 hours
Hosts and 404s
Redirect www to perma.sh, canonical tag, real 404 status.
Founding Engineer 2 hours
Submit
Google Search Console and Bing Webmaster Tools.
Growth Marketer 1 hour
Check previews
Paste the link in Slack, WhatsApp, LinkedIn and X.
Growth Marketer 1 hour
Why it works
Build the landing page to static HTML at deploy time (Vite has pre-render plugins, or render Page.jsx once with react-dom/server) so the full English copy is in the HTML. Fix the noscript text per page, add robots.txt and a sitemap with the homepage, legal pages and future pages, add an og:image (the vault illustration with "€1/GB, paid once"), redirect www to the apex, add a canonical tag, and return a real 404 for unknown paths.
A growth hire can only compound search, AI and word of mouth if the page those channels read has words on it. It is a few days of engineering and every later idea depends on it.
Ready to use
ChecklistCrawlability fix list
[ ] noscript text describes perma.sh (not legal documents) [ ] Homepage copy present in served HTML [ ] /robots.txt with sitemap line [ ] /sitemap.xml with every real page [ ] og:image and twitter:image (1200x630) [ ] www.perma.sh 301 to perma.sh [ ] <link rel="canonical"> on each page [ ] Unknown paths return 404 [ ] Search Console and Bing verified
Risks and how to handle them
Pre-rendered and live copy drift apart
Generate the HTML from the same component at build time.
Language switch breaks
Serve the Spanish version at its own URL (idea 6).
Evidence
- Your websiteThe HTML perma.sh serves (checked November 4, 2026) is an empty React shell: a title, a meta description ("Client-side encrypted, write-once storage for businesses and individuals. €1/GB, paid once. No subscription.") and an empty root element. The only readable body text is a noscript message: "These legal documents require JavaScript. Email [email protected] for a copy." The same shell, with the same message, is served on the homepage and every other path.Source
- Your websiterobots.txt and sitemap.xml both return 404. www.perma.sh serves the same page without redirecting to perma.sh, and there is no canonical tag. Any made-up path (for example /pricing or /this-page-does-not-exist) returns status 200 with the homepage shell. The page sets a large-image Twitter card but has no og:image or twitter:image, so shared links show no picture.Source
- Google PageSpeedMobile performance 100/100 (desktop 100): content visible in about 1.0 s, 317 KB in 4 requests, no third-party code. One accessibility issue: low colour contrast.Source