See llms.txt for all machine-readable content.

Back to Templates

Draft beginner-friendly AI Instagram post drafts with OpenRouter and Google Sheets

Created by

Created by: Jun Kamiya || ac154602
Jun Kamiya

Last update

Last update 2 days ago

Categories

Quick overview

This daily workflow pulls AI news from multiple RSS feeds, uses OpenRouter LLMs to score articles for beginner usefulness, drafts Japanese Instagram captions and carousel slides, then saves the drafts to Google Sheets and emails a digest via Gmail for human review.

How it works

  1. Runs every day at 08:00 (Asia/Tokyo) or manually when triggered.
  2. Reads enabled RSS feeds, normalizes article metadata, removes old items, and deduplicates entries across feeds.
  3. Loads the existing “Content Queue” from Google Sheets and filters out articles whose normalized URLs already exist, keeping only a limited set for scoring.
  4. Sends the remaining article titles and summaries to an OpenRouter chat model to score each one for AI-beginner utility, beginner fit, actionability, reliability, and freshness, and to assign a category and Japanese rationale.
  5. Ranks the scored items using a weighted formula and selects up to the configured number of high-quality candidates, with a soft preference for category variety.
  6. For each selected article, uses an OpenRouter writing model to generate a Japanese Instagram post draft (caption, 5–7 carousel slides, CTA, hashtags, and unverified points).
  7. Flags potential risks (for example, overly absolute wording, pricing/availability claims, or missing source details), writes/updates the draft rows in Google Sheets, and emails an HTML digest of the candidates via Gmail.

Setup

  1. Add an OpenRouter credential and connect it to both OpenRouter chat model nodes used for scoring and writing.
  2. Add a Google Sheets credential, select your spreadsheet, and create a “Content Queue” worksheet with the required header columns before running.
  3. Add a Gmail credential and set the recipient email address in the email step.
  4. Review and edit the RSS feed list in the feed-configuration code step to match your preferred sources.
  5. Adjust the workflow settings (age window, maximum articles to score, candidate limits, minimum score, and slide count) to fit your editorial process.