Work

AI editorial pipeline

SGNews

I built SGNews's AI editorial pipeline, which rewrites source stories into short briefings every five minutes, and repaired how it collects stories and detects failures.

SGNews editor with a source article beside a local story draft, duplicate checks and publishing controls.
47,277articles publishedPublic WordPress API count, 7 May–12 September 2026.
Client
Digital news
Role
Built
Timeline
2026 · Delivered
Stack
Python, Claude, WordPress REST API, MySQL, PHP, cPanel/WHM, cron

Just reviewed the new articles on staging and honestly the quality difference is night and day. The key points are now pulling real quotes, specific names, dates, and data that genuinely add value beyond the summary.

Juan Minatta, Founder, SGNewsDigital news

Every five minutes, AI rewrites new source stories and the pipeline checks and publishes them. Problems with individual sources can be identified and corrected.

Articles were rewritten by hand, and hidden plugin flags made scheduled feed runs skip processing. Feeds that looked healthy could quietly stop producing drafts.

  1. Built an AI step that rewrites each source story as a new headline, short summary, and key points, with a link to the source.
  2. Added checks for duplicate stories and Singapore relevance.
  3. Investigated news sources that stopped producing articles and repaired how their content was collected.

Keep articles moving

Process the next waiting articles instead of repeatedly reading the same first page.

Show why a source failed

Distinguish a missing article, an empty page, and a blocked source so the problem can be investigated.

What the project delivered

2,445articles in one measured weekPublic WordPress API, 6–12 September 2026.
89 secondsto clear the backlog13 previously skipped articles processed with zero errors.

Say hello.

Tell me what you'd like to build or improve.

Quick navigation

Find a page.

Use the arrow keys to choose a result, Enter to open it, or Escape to close.