SEO and migration
Bring the old site with you,and leave its bad habits behind.
castomUI crawls the website you already have, lists every page it finds, and lets you scrape, analyse and rebuild them one at a time or all at once. Then it gives you one table for every title and description in the project, the switches search engines actually read, and a public description written for the language models that now read your site too.
The web crawler:
start from what exists.
Point the crawler at a website and choose how deep it should follow links. What comes back is a list of found pages with a checkbox beside each, which is the first honest inventory most old sites have had in years.
For any page on that list you can scrape it — meta title, description and keywords, the content itself and its OpenGraph tags — analyse it with AI, and add it to this project's sitemap as a new page. When the list is long, the same three moves run in bulk: scrape all, analyse the selected pages, prepare the selected pages.
The web checker is the single-page version: give it one address and it tells you what that page is, its properties and its content. Useful for a competitor's landing page, and for the one page on your old site nobody can explain.
- ✓/Scraped · analysed · added to sitemap
- ✓/about-usScraped · analysed · ready to prepare
- ✓/services/installationScraped · waiting for analysis
- ○/news/2019-summer-hoursFound · left unticked on purpose
- !BulkScrape all · Analyse selected · Prepare selected
Example addresses, not a real siteNothing is added until you add it
Per page, and for all of them
Every signal a search engine reads,
in one place.
Meta tags per page
Title, description and keywords on every page, written by hand or generated with AI from the page's own content, its description and the purpose you gave it.
The META tab
All pages in one table, with URL, title, description, keywords and canonical editable inline. Two buttons do the heavy lifting: Analyze all pages and Generate META for all pages.
Canonical address
Set per page, so the product that lives at three URLs is counted once, at the address you meant.
Index and follow
Index or noindex, follow or nofollow, page by page. The thank-you page and the internal price list stay out of results without a line of code.
Sitemap control
The sitemap is built from your page tree. Mark a page "not in sitemap", or hide it completely; drafts never appear because drafts never publish.
Redirection list
301 for moved for good, 302 for moved for now. The old URLs a crawl turned up point at their new homes, and years of inbound links keep working.
Every page of the old site,
accounted for before the new one goes live.
How the site looks when it is shared.
OpenGraph defaults are set once for the project: site name, title, description, an image from the media library or a URL, and the Twitter site and author accounts. The type is chosen from the ones the protocol knows — website, article, product, profile, book, video, music, business, place.
Any page can override the defaults with its own OpenGraph, so a product page shares as a product and an article shares as an article, rather than every link in a chat looking like your homepage.
- PProject defaultType: business · Image: from media library · Twitter site set
- 1/journal/choosing-oakOverride · Type: article · Own title and image
- 2/shop/oak-tableOverride · Type: product · Own description
- =Everything elseInherits the project default
Invented pagesPages override, never the reverse
llms.txt:
a description for the machines that read you.
Search is no longer only a list of links. Language models summarise sites for people who never click through, and they do better with a plain statement of what you are. castomUI publishes a public description for LLMs as llms.txt, and can generate it with AI from what the project already knows.
It sits beside the private project description, which is different on purpose: the private one is context that makes castomUI's own AI output better and is never published; the public one is written to be read by strangers.
On keywords: the field is there, and castomUI will fill it in for you. Whether any search engine still reads it is a question that has outlived several careers in SEO.
Speed and polish
The small things that count as quality.
Prefetch on hover
Subpages start loading while the pointer is still on the link, so the click feels instant. One switch in the project's advanced settings.
Service worker
Pages served fast on return visits and still available when the connection drops. Also one switch.
Web App Manifest
A full name and a short name, and the site can be installed on a phone's home screen like an app.
Favicons
A 16×16 .ico and a 64×64 .png, uploaded once in project settings, so the tab and the bookmark both look like you.
Questions
The ones people ask before moving a site.
Does the crawler copy the old design?
No. It brings across content, meta tags and OpenGraph, and the AI analysis helps you decide what each page should become. The new pages are built from your components, which is usually the reason you are moving.
Will old links break?
Not if they are in the redirection list. Add a 301 for every URL that changes, and inbound links and bookmarks land on the new page.
Can I review AI-generated META before it publishes?
Yes. Generated titles and descriptions land in the META tab, where every field is editable inline, and nothing reaches the live site until you publish.
Can I keep a page live but out of search?
Yes — set noindex on the page, take it out of the sitemap, or both. Hiding it completely is a separate switch.
Every window,
one website.
Open the editor and arrange them the way the job needs.
Open castomUI See pricing