# An existing corporate site rebuilt page for page

Entry B03. Sector: Engineering firm. Status: in production. Delivered: 2026-06.
Source: https://ai.prospicience.in/work/site-rebuilt-from-a-scrape

For an engineering firm, we rebuilt a working corporate site on a modern stack by first extracting every page's content into structured data, then rebuilding against it. The result is a modern, more secure site with the same content, checked page for page against the original. The first full rebuild took a day.

Speed: First full rebuild in a day.

## The challenge

A working corporate site needed a modern stack, but a manual rebuild risks dropped pages, missing product detail and broken structure that nobody notices until a customer does.

## Why it mattered

The site carries the firm's product and engineering content. Losing any of it in a move means lost enquiries and lost search visibility, so the migration had to be provable, not just plausible.

## What we built

- Extracted the existing site's content into structured data before any rebuilding started.
- Rebuilt every page from that data, adding real product images and detailed engineering content.
- Compared old against new, page for page.
- Added security headers and a spam protected, accessible contact form.

## The result

- The same content on a modern stack, with nothing lost in transcription.
- Security headers in place, and a contact form with accessibility fixes and spam protection.
- A sitemap and search settings set up for the firm's own domain.

## Built for trust

Every page is traceable to its original. Content was extracted to structured data rather than copied by hand, so each rebuilt page could be compared against the source instead of taken on trust.

## Automation at work

The old site's content now lives as a structured record that stays with the firm, and every rebuilt page is checked against it. A sitemap and search settings for the firm's own domain tell search engines exactly what to read.

What it produces: a structured record of every page, page for page comparisons against the original, a sitemap for search engines, a spam protected contact form.

- A content change starts from data a machine can read and compare, not from someone retyping a page.
- The next check or move reuses the same record, so it starts where this one finished.
- Spam protection keeps the team's inbox for real enquiries.

How it works day to day: Content was pulled from the original site into structured data, the new pages were built from it, and each was compared back against the source before launch. People reviewed the comparison and signed off the move.

## Figures from the delivered system

- approach: content extracted, then rebuilt
- comparison: page for page

## The technology

- Same content, modern stack
- Nothing lost in the move
- Old and new compared side by side

### Technical notes

- Content extraction to structured JSON as the migration source of truth
- Static generation from the extracted content

## Questions

### How do you migrate a site without losing content?

Extract first, rebuild second. Pulling every page into structured data creates a fixed record of the original, and each new page is built from it and compared back against it.

### How quickly was it rebuilt?

The project history shows the first full rebuild, including real product images and engineering content, completed on the day the project started. Security hardening and search work followed.

Capabilities: platform-migration.