I built a fully automated GitHub trending repo discovery and reporting system in n8n, powered by Google Gemini for AI summarization and translation.

The workflow runs on a schedule, pulling the last 30 days of push history and the last 24 hours of trending repos, then filters out repos already posted and calculates the top candidates. For each qualifying repo, it fetches the README and passes it to Gemini to generate a clean summary. That summary is merged into Markdown, converted to HTML, and emailed out automatically — while a parallel branch uses AI to generate a repo description, translate it to English, and push both the original and English versions as files directly into GitHub. Every posted repo is logged to prevent duplicate reporting on future runs.

This isn't a simple trigger-and-notify setup — it's a branching pipeline with AI decision-making at three separate points (summarization, description generation, and translation), all coordinated so nothing gets posted twice and everything stays properly documented across formats: email report, Markdown file, and multilingual GitHub files.