Search engines see the same page as visitors
The site is an Angular application, but crawlers get a server-rendered page with the same title, description, text and structured data, and pages that are switched off answer 404.
The problem
A single-page application sends an almost empty HTML page and fills it with JavaScript. Some crawlers and most link previews (LinkedIn, Slack, messengers) never run that JavaScript, so they see no title, no text and no image.
How dok.sh solves it
- One source for page metadata. For every URL in both languages the backend decides the title, description, canonical address, hreflang links, preview image and JSON-LD. The Angular app and the crawler version both use it.
- Crawlers get real HTML. nginx recognises search engines, AI crawlers and link-preview bots by their user agent and sends them a server-rendered page with the same heading and text as the app.
- One linked JSON-LD graph: the person, the consulting service, the website, and per page a breadcrumb, a list of services or an article for case studies and use cases.
- Honest status codes. An unknown address or a section switched off in the admin returns 404 with
noindex. A list page with nothing in it is kept out of the sitemap.

The result
A link shared on LinkedIn shows the right title and picture, and search engines index the same text people read.
Technologies
SEO, nginx, Spring Boot, JSON-LD
- Published: 2026-11-03
- Project: dok.sh — Portfolio and CV Platform
- Case study