<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Shoki | Kaizen Automation</title>
    <description>The latest articles on DEV Community by Shoki | Kaizen Automation (@kaizenautomation).</description>
    <link>https://dev.to/kaizenautomation</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3967564%2Ff0d3577a-c07e-4208-bb20-a0179eec72ba.png</url>
      <title>DEV Community: Shoki | Kaizen Automation</title>
      <link>https://dev.to/kaizenautomation</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/kaizenautomation"/>
    <language>en</language>
    <item>
      <title>The n8n Workflow I Built to Auto-Translate Japanese AI News Into English Every Morning ($0.03/run)</title>
      <dc:creator>Shoki | Kaizen Automation</dc:creator>
      <pubDate>Thu, 04 Jun 2026 05:51:13 +0000</pubDate>
      <link>https://dev.to/kaizenautomation/the-n8n-workflow-i-built-to-auto-translate-japanese-ai-news-into-english-every-morning-003run-1po6</link>
      <guid>https://dev.to/kaizenautomation/the-n8n-workflow-i-built-to-auto-translate-japanese-ai-news-into-english-every-morning-003run-1po6</guid>
      <description>&lt;p&gt;Japan's AI scene moves fast — but almost everything is in Japanese.&lt;/p&gt;

&lt;p&gt;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: &lt;strong&gt;$0.03 per run&lt;/strong&gt;.&lt;/p&gt;

&lt;p&gt;Here's exactly how I built it.&lt;/p&gt;

&lt;h2&gt;
  
  
  The Problem
&lt;/h2&gt;

&lt;p&gt;I follow Japanese AI research closely, but manually translating 10+ articles every morning was killing my time. I needed automation.&lt;/p&gt;

&lt;h2&gt;
  
  
  The Workflow Architecture
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;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)
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Key Implementation Details
&lt;/h2&gt;

&lt;h3&gt;
  
  
  1. The AI Translation Node
&lt;/h3&gt;

&lt;p&gt;I use Claude's API via HTTP Request node. The prompt is structured to return JSON:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"title_en"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"..."&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"summary_en"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"..."&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"importance_score"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="mi"&gt;8&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"tags"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"LLM"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"robotics"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  2. Cost Control (Critical)
&lt;/h3&gt;

&lt;p&gt;Before hitting the AI node, I added an IF filter:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Only process articles from the last 24 hours&lt;/li&gt;
&lt;li&gt;Max 15 articles per run&lt;/li&gt;
&lt;li&gt;Cache results in Google Sheets to skip duplicates&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;This cut my Claude API costs by ~60%.&lt;/p&gt;

&lt;h3&gt;
  
  
  3. Error Handling
&lt;/h3&gt;

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

&lt;h2&gt;
  
  
  The Result
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Cost&lt;/strong&gt;: $0.03/run ($0.90/month)&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Time saved&lt;/strong&gt;: ~45 minutes daily&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Articles processed&lt;/strong&gt;: 10-15 per morning&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Reliability&lt;/strong&gt;: 47 days of uninterrupted runs&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Workflow File
&lt;/h2&gt;

&lt;p&gt;The full n8n workflow JSON is available in my &lt;a href="https://shokiwork.gumroad.com/l/vazfgc" rel="noopener noreferrer"&gt;Gumroad starter pack&lt;/a&gt; — it's part of the 5-workflow bundle for $9.&lt;/p&gt;

&lt;h2&gt;
  
  
  What's Next
&lt;/h2&gt;

&lt;p&gt;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.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;This article was originally published on &lt;a href="https://kaizenautomation.substack.com/p/the-n8n-workflow-i-built-to-auto-translate-japanese-ai-news" rel="noopener noreferrer"&gt;Kaizen Automation&lt;/a&gt; — weekly n8n &amp;amp; AI automation workflows, filtered through Japanese precision.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>tutorial</category>
    </item>
  </channel>
</rss>
