The n8n Workflow I Built to Auto-Translate Japanese AI News Into English Every Morning ($0.03/run)

Japan's AI scene moves fast — but almost everything is in Japanese.

I built an n8n workflow that runs every morning at 7 AM, scrapes the top Japanese AI news, translates each article using Claude AI, and delivers a clean English briefing to my inbox. Total cost: $0.03 per run.

Here's exactly how I built it.

I follow Japanese AI research closely, but manually translating 10+ articles every morning was killing my time. I needed automation.

Schedule Trigger (7 AM daily)
→ HTTP Request (fetch RSS feeds from Japanese AI sources)
→ Split in Batches (process each article)
→ IF node (filter: published today only)
→ HTTP Request (Claude API - translate + summarize)
→ Code node (parse JSON response, format output)
→ Gmail (send ranked daily briefing)

I use Claude's API via HTTP Request node. The prompt is structured to return JSON:

{
  "title_en": "...",
  "summary_en": "...",
  "importance_score": 8,
  "tags": ["LLM", "robotics"]
}

Before hitting the AI node, I added an IF filter:

This cut my Claude API costs by ~60%.

Each source runs through its own Try/Catch node. If one RSS feed fails, the others continue. The Gmail summary flags partial runs automatically.

The full n8n workflow JSON is available in my Gumroad starter pack — it's part of the 5-workflow bundle for $9.

I'm building a v2 that scores articles by relevance to my specific interests (AI agents, automation tools) and only sends the top 5. Fewer emails, higher signal.

This article was originally published on Kaizen Automation — weekly n8n & AI automation workflows, filtered through Japanese precision.