← back to work

Automated content research pipeline that surfaces genuinely viral Reels

A content research initiative · Media / content operations · Viral content discovery and analysis

Finding genuinely viral content meant more than tracking view counts — it meant spotting mid-size creators having breakout moments before those moments went stale. That kind of analysis was taking eight-plus hours of manual research per cycle. This pipeline does it twice a day, automatically.

The results

  • 15 hours of manual research eliminated per week
  • 5–15 genuinely viral posts surfaced per run, each with an AI-written breakdown of hook, structure, and content angle
  • Runs unattended, twice daily, seven days a week
  • Catches breakout creators that raw-follower-count tools miss — the system rewards relative performance, not audience size
n8n workflow showing the full Instagram viral content pipeline — schedule trigger, Apify Explore scrape, outlier detection, Whisper transcription, Gemini analysis, and Google Sheets output
n8n workflow — schedule to Sheets, 15 nodes
Instagram Viral Content Pipeline ArchitectureFive stages in a single left-to-right flow. Explore Feed is scraped every 12 hours. Creator Filter removes duplicate accounts. Outlier Detection flags posts at 2.5 times a creator's baseline view count. AI Breakdown transcribes audio with Whisper and analyzes with Gemini. Results append to a Research Sheet in Google Sheets.Explore FeedScraped every 12hCreator FilterUnique accounts onlyOutlier Detection2.5× their baselineAI BreakdownWhisper + GeminiResearch SheetGoogle Sheets

Tech stack

  • n8n
  • Apify
  • OpenAI Whisper
  • Gemini 2.5 Flash
  • Google Sheets