# A 30 part technical catalogue that reads without JavaScript

Entry B04. Sector: Forging manufacturer. Status: in production. Delivered: 2026-06.
Source: https://ai.prospicience.in/work/technical-catalogue-site

For a forging manufacturer, we built a 30 part technical catalogue with drawings to a published standard and a quote request on every part. It reads fully on a phone with a weak connection before any script loads, so an engineering buyer on a plant floor always gets the specification. Built in a week.

Speed: Built in a week.

## The challenge

Engineering buyers judge a supplier on specifications, standards and drawings, often from a plant floor on a weak connection. A catalogue that fails to load there loses the enquiry.

## Why it mattered

For a quote driven manufacturer, the website is where a buyer decides whether to send a drawing. Every part that cannot be inspected is a quote that never arrives.

## What we built

- Catalogued 30 parts with technical drawings to a published standard.
- Added a quote request path from every part page.
- Built every page to read in full before any script loads.
- Cut image weight from 63 MB to 2 MB and met the WCAG 2.1 AA accessibility standard.

## The result

- A complete catalogue buyers can inspect on any device and any connection.
- A path to a quote from every one of the 30 parts.
- Images cut from 63 MB to 2 MB, so pages load quickly on mobile data.
- A recognised accessibility standard met, with a visible focus treatment for keyboard users.

## Built for trust

Every part page is readable with scripts switched off. The specification always reaches the buyer, because a buyer looking for a part was already looking for this supplier.

## AI at work

Each of the 30 parts is described to search engines and AI assistants as a product from a named manufacturer, so an engineer who asks an assistant for a brass or copper part can be pointed straight to the right drawing and a quote request.

What it produces: a structured product listing for each of the 30 parts, a machine readable profile of the manufacturer and its location, page trails search engines follow, a quote request path from every part.

- Engineering buyers who research with AI assistants find the specification, not just the company name.
- A new part added to the catalogue gets the same structured listing automatically, so the catalogue grows without extra search work.
- Images cut from 63 MB to 2 MB keep every page quick to read and quick to index.

How it works day to day: The catalogue build writes a structured description for every part next to its visible specification. Search engines and AI assistants read it when they index the site, and the team adds parts while the description follows.

## Figures from the delivered system

- parts catalogued: 30
- readable without JavaScript: every page
- accessibility standard: WCAG 2.1 AA
- image weight: 63 MB cut to 2 MB

## The technology

- Thirty parts with real technical drawings
- Request a quote from any part
- Works on a weak connection

### Technical notes

- Progressive enhancement with server rendered content as the baseline
- Static generation with a per part quote route
- WCAG 2.1 AA with a visible focus treatment

## Questions

### Why build a catalogue that works without JavaScript?

Buyers often check parts from a factory floor on a weak connection. Content that arrives before any script means the drawing and specification always get through.

### How long did the catalogue take?

The project history shows the first build, the real drawings, and the performance and accessibility pass all completed within a week.

### Can buyers request a quote online?

Yes. Every one of the 30 part pages has its own quote request path.

Capabilities: platform-migration, document-intelligence.