AI & Automation Stack · FreeMalta

Web Intelligence Stack

Q: What is the difference between SEO, AEO and GEO?

SEO (Search Engine Optimisation) is optimising for traditional search engines like Google — getting pages to rank for specific keywords. AEO (Answer Engine Optimisation) is optimising for AI-powered answer engines like Perplexity, ChatGPT Search and Google AI Overviews — these extract and synthesise answers, and cite sources. GEO (Generative Engine Optimisation) is optimising for Large Language Models that generate responses — making sure your content is trusted, structured and citable by AI systems. In 2026, a serious web strategy needs all three.

Q: What is programmatic SEO and how does FreeMalta use it?

Programmatic SEO is creating large numbers of unique, data-driven pages at scale — each targeting a specific search query. FreeMalta uses this across every property: the Pharmacy tool generates pages for every locality and pharmacy in Malta (400+ pages), Markets and Garage cover company profiles and categories (300+ pages), and News Beast publishes 27 new indexed articles every day. Each page has its own URL, metadata and schema — generated automatically from structured data.

Q: What is IndexNow and why does it matter for News Beast?

IndexNow is a protocol that allows websites to instantly notify search engines when content is published — rather than waiting for a crawler to discover it. News Beast publishes 27 articles per day. Without IndexNow, Google might take days to discover new articles. With IndexNow, each article is submitted to search engines within seconds of publication. FreeMalta articles typically appear in Google Search Console within 2–4 hours of being published.

Q: What is llms.txt and why does it matter in 2026?

llms.txt is an emerging standard that tells Large Language Models what content on your site is available and how it's structured. As AI answer engines like Perplexity, ChatGPT and Claude increasingly serve answers by reading and citing web content, a well-structured llms.txt ensures your site is correctly understood and cited. FreeMalta's 2,000+ indexed pages contain unique Malta-specific data that AI answer engines increasingly surface when users ask Malta-related questions. The citation is the new backlink.

Q: What is MCP and how does it relate to web intelligence?

MCP (Model Context Protocol) is Anthropic's open standard for connecting AI models to external data sources and APIs. In a web intelligence context, MCP allows AI agents to directly query your database or call your APIs — without scraping your website. For FreeMalta, this means AI integrations can pull live Malta data directly into an AI workflow via MCP, returning live structured data rather than cached web content.

Q: What schema markup does FreeMalta use?

FreeMalta implements six schema types across its properties: BreadcrumbList (every hub and partner page), FAQPage (all content pages — heavily used by AI answer engines for citation), Organization (brand identity and trust signals), NewsArticle (every News Beast article — eligible for Top Stories and faster indexing), LocalBusiness/Pharmacy (pharmacy and local business pages with opening hours and ratings), and AggregateRating (company profiles in Markets and Garage). Schema markup is implemented automatically via PHP templates — not manually per page.

A beautiful website nobody finds is an expensive business card. In 2026, visibility means three things: ranking on Google (SEO), being cited by AI answer engines (AEO), and being trusted by LLMs (GEO). Here is how FreeMalta built for all three.

Collect

Firecrawl

Web data

Research

Perplexity

AI search

Process

N8N

Orchestrate

or Make.com

Deploy

Emergent

AI agents

Serve

Durable

Frontend

2,000+

Indexed pages across FreeMalta

News Beast articles per day

400+

Pharmacy pages (localities × pharmacies)

300+

Markets & Garage pages (companies + categories)

Firecrawl

Web Data Collection

Collect

How FreeMalta uses it — News Beast

News Beast needs to monitor 8+ Malta news sources, scrape new articles every 2 hours, and pass clean content to Claude API for categorisation and rewriting.

Firecrawl runs on a Cloudflare cron trigger. Every 2 hours it scrapes each source and returns clean markdown — no HTML noise, no JavaScript rendering issues, no bot detection. Claude API receives content it can actually read. The News Beast pipeline runs without maintenance.

Works well when

Feeding clean data directly to Claude API

Monitoring competitor or news sites at scale

Enriching company profiles from websites

High-frequency scraping without maintenance

Watch out

Charges per page crawled — budget your usage

Some sites actively block crawlers

Not for real-time data (30s+ latency)

Get Firecrawl Partner page

Claude API + Supabase + Cloudflare

Process · Store · Serve at the Edge

Core Stack

The News Beast architecture

Firecrawl returns an article in markdown. Claude API categorises it (iGaming / Politics / Business / etc.), rewrites the headline in editorial voice, generates a 150-word summary, creates an SEO-optimised slug, and returns structured JSON — all in one API call.

Supabase D1 (SQLite at Cloudflare's edge) stores the article. Cloudflare Worker serves requests from the nearest edge node — Frankfurt for Malta readers. No origin server hit for cached content. TTFB: ~80ms.

The full pipeline costs under €20/month at current News Beast traffic. Scale 10x — still under €50/month. This is the architecture that makes programmatic SEO economically viable.

Full pipeline

Firecrawl (scrape) → Claude API (categorise + rewrite + slug) →
Supabase D1 (store at edge) → Cloudflare KV (cache) →
Cloudflare Worker (serve · 80ms TTFB) → IndexNow (submit to Google)

NewsArticle schema is injected automatically per article. IndexNow fires on every new insert. The dynamic sitemap rebuilds from D1 on every request. Google sees new content within 2–4 hours of publication.

FreeMalta architecture

Programmatic SEO — how 2,000+ pages get built and indexed

Programmatic SEO is the practice of creating large numbers of unique, data-driven pages at scale — each targeting a specific search query. The key word is unique. Not duplicated content with one word changed. Genuinely different pages, each with its own data, schema and metadata.

Property	Template logic	Example slug	Scale
Pharmacy Tool	1 template × locality × pharmacy × opening hours	`/pharmacies/sliema`	400+ pages
Markets / Garage	1 template × company × sector + category pages	`/markets/company-stories/acme-igaming`	300+ pages
News Beast	1 template × article × category × date	`/news/igaming/article-slug`	27/day · growing
Hub Partners	1 template × partner × category	`/hub/partners/deel`	44 partner pages
Hub Stacks	1 template × stack × FAQ schema	`/hub/outbound-sales`	8 stack pages

Every page type has a matching schema implementation — Local Business + Rating for pharmacy pages, NewsArticle for News Beast, Organization + FAQ for hub pages. Schema is injected via PHP template, not manually per page. Write the template once. Generate the schema a thousand times.

Structured data

Schema markup — what FreeMalta implements and why

BreadcrumbList

Shows page hierarchy in Google results. Users know where they are before clicking. Reduces bounce rate.

Used on: Every hub page, partner page, stack page

FAQPage

Surfaces Q&A in Google results. Also heavily used by AI answer engines to extract structured answers for citations.

Used on: All hub, partner and stack pages

NewsArticle

Eligible for Top Stories carousel and News tab. Critical for News Beast indexing speed and Google News visibility.

Used on: Every News Beast article

LocalBusiness + Pharmacy

Powers rich results for local searches — opening hours, location, on-duty status. Core to the pharmacy tool's search visibility.

Used on: All 400+ pharmacy pages

AggregateRating

Displays star ratings in search results. Increases CTR by 15–30%. Applied to company profiles in Markets and Garage.

Used on: Company profiles, Garage listings

Organization

Establishes entity identity for Google's Knowledge Graph. Critical for brand searches and AI model trust signals.

Used on: About page, partner detail pages

Indexing speed

IndexNow + Dynamic Sitemap — indexed in hours, not days

News Beast publishes 27 articles per day. Without proactive indexing, Google's crawler might take days to discover new content. Indexing speed is a competitive advantage in news.

indexnow-worker.js · fires on every article insert

# Article stored → IndexNow submits URL immediately

→ await fetch('https://api.indexnow.org/indexnow', {

body: JSON.stringify({ host: 'freemalta.com', urlList: [articleUrl] })

})

✓ Submitted to Bing + Yandex (+ Google via partnership) in <1s

# Dynamic sitemap — always current, no manual updates

→ GET /sitemap-news.xml

→ SELECT slug, category, generated_at FROM articles ORDER BY generated_at DESC

✓ Sitemap reflects current content · regenerated on every request from D1

The AI frontier — GEO

llms.txt and MCP — being found by AI, not just Google

In 2026, an increasing share of information retrieval happens not through Google, but through AI interfaces. Users ask Perplexity, ChatGPT or Claude a question and get a synthesised answer with citations. The sites that get cited are the new page 1.

llms.txt is an emerging standard (analogous to robots.txt) that tells LLMs what content on your site is available and how it's structured. MCP (Model Context Protocol) — Anthropic's open standard — goes further: it allows AI agents to directly query your database or APIs without scraping your website. FreeMalta's live data could be surfaced by a Claude agent via MCP, returning current figures rather than cached web content.

FreeMalta's 2,000+ indexed pages of unique Malta-specific data are exactly what AI answer engines need — specific, verifiable, regularly updated, with clear authorship. The citation is the new backlink.

FreeMalta builds what it recommends.

News Beast. AnotherCandidate. This platform.
AI agents. Automated newsrooms. Workflows that run while you sleep.

No templates. No "I came across your profile." No decks full of buzzwords.

If you need it built — let's talk.

Fractional CAIO

Off The Record

How News Beast Gets 27 Articles Indexed on Google Every DaySoon SEO vs AEO vs GEO: What Actually Matters for Malta Businesses in 2026Soon Programmatic SEO: The Architecture Behind FreeMalta's 2,000 PagesSoon llms.txt: The File That Could Put Your Site in Every AI AnswerSoon

Frequently Asked Questions

What is the difference between SEO, AEO and GEO?

SEO (Search Engine Optimisation) is optimising for traditional search engines like Google — getting pages to rank for specific keywords. AEO (Answer Engine Optimisation) is optimising for AI-powered answer engines like Perplexity, ChatGPT Search and Google AI Overviews — these extract and synthesise answers, and cite sources. GEO (Generative Engine Optimisation) is optimising for Large Language Models that generate responses — making sure your content is trusted, structured and citable by AI systems. In 2026, a serious web strategy needs all three.

What is programmatic SEO and how does FreeMalta use it?

Programmatic SEO is creating large numbers of unique, data-driven pages at scale — each targeting a specific search query. FreeMalta uses this across every property: the Pharmacy tool generates pages for every locality and pharmacy in Malta (400+ pages), Markets and Garage cover company profiles and categories (300+ pages), and News Beast publishes 27 new indexed articles every day. Each page has its own URL, metadata and schema — generated automatically from structured data.

What is IndexNow and why does it matter for News Beast?

IndexNow is a protocol that allows websites to instantly notify search engines when content is published — rather than waiting for a crawler to discover it. News Beast publishes 27 articles per day. Without IndexNow, Google might take days to discover new articles. With IndexNow, each article is submitted to search engines within seconds of publication. FreeMalta articles typically appear in Google Search Console within 2–4 hours of being published.

What is llms.txt and why does it matter in 2026?

llms.txt is an emerging standard that tells Large Language Models what content on your site is available and how it's structured. As AI answer engines like Perplexity, ChatGPT and Claude increasingly serve answers by reading and citing web content, a well-structured llms.txt ensures your site is correctly understood and cited. FreeMalta's 2,000+ indexed pages contain unique Malta-specific data that AI answer engines increasingly surface when users ask Malta-related questions. The citation is the new backlink.

What is MCP and how does it relate to web intelligence?

MCP (Model Context Protocol) is Anthropic's open standard for connecting AI models to external data sources and APIs. In a web intelligence context, MCP allows AI agents to directly query your database or call your APIs — without scraping your website. For FreeMalta, this means AI integrations can pull live Malta data directly into an AI workflow via MCP, returning live structured data rather than cached web content.

What schema markup does FreeMalta use?

FreeMalta implements six schema types across its properties: BreadcrumbList (every hub and partner page), FAQPage (all content pages — heavily used by AI answer engines for citation), Organization (brand identity and trust signals), NewsArticle (every News Beast article — eligible for Top Stories and faster indexing), LocalBusiness/Pharmacy (pharmacy and local business pages with opening hours and ratings), and AggregateRating (company profiles in Markets and Garage). Schema markup is implemented automatically via PHP templates — not manually per page.