Automated AI News Pipeline (this site)

The pipeline behind this site’s Tech News: scheduled GitHub Actions scrape sources, an LLM cascade on Groq rewrites and enhances articles, and layered quality gates (date integrity, language checks, instruction-leak detection, duplicate detection) decide what gets published, with Telegram alerts.

Runs on a schedule with preflight cost checks · Model cascade with fallback tiers

Problem

Publishing AI-generated content safely needs more than a prompt: models fail, sources change and output can leak instructions.

Approach

• Preflight step skips paid API calls when there is nothing new.
• Model cascade with fallback and last-resort tiers, verified against the live Groq model list before each run.
• Quality gates reject leaked instructions, wrong-language output, bad dates and duplicates before anything is saved.

Outcome

• Fully automated publishing with Telegram reporting

Stack

Node.js · Groq · Firecrawl · Supabase · GitHub Actions · Telegram Bot API · Vercel

Links

• See it live

Related skills

• RAG & Grounded LLM Systems
• Agentic & LLM Automation

Work together

Open to AI engineering roles in Dublin, across Ireland and remote, and to freelance projects. Get in touch.