Automate Instagram Stories with AI (Make.com & n8n)
This tutorial demonstrates how to automate Instagram Stories using Make.com and n8n, with AI-generated images via Replicate and posting through Blotato. Two versions are shown: a simple direct-publish workflow and an advanced version with human review via Airtable. The host credits consistent story posting with achieving his largest revenue month ever.
Summary
The video features a host and AI automation expert Kevin walking through two different automation systems for posting Instagram Stories without manual daily effort. The core stack involves Replicate.com for AI image generation (using models like Flux or Imagen 3/Nano Banana), Airtable for content review and status tracking, and Blotato as the Instagram publishing API.
The first and simpler version ('Option A') in Make.com consists of just three modules: a prompt input node, a Replicate image generation call using the Flux model, and a direct Blotato post to Instagram Stories. Users only need to set their image prompt, call to action, and API keys. This version runs without human review and publishes immediately. A typo in the generated output during the live demo highlights one limitation of this approach.
The second, more advanced version introduces a human-in-the-loop via Airtable. An AI agent (using OpenAI) takes the user's prompt and call to action and generates three distinct variations with different styles, moods, or compositions. Replicate then generates all three images, and each is saved as a record in Airtable with a 'To Review' status. The user reviews the three options and changes the preferred one to 'Approved.' A second workflow (or trigger in n8n) scans for approved records, publishes the story, and updates the status to 'Published' to prevent re-posting.
A key technical difference between Make.com and n8n is noted: Make.com requires two separate blueprint files since it cannot have multiple entry points in one workflow, while n8n supports multiple triggers within a single workflow, making it easier to manage both parts together.
The host shares that posting stories consistently for 10 days resulted in his largest revenue month, emphasizing that story viewers are self-selected engaged followers. Kevin recommends starting with a clear goal—what you're selling and what action you want viewers to take—as the foundation for effective prompts. The recommended posting frequency for beginners is three to four times per week. Templates for both Make.com and n8n versions are made available in the video description.
Key Insights
- The host claims that posting Instagram Stories consistently for 10 days using this AI automation resulted in his largest revenue month ever, attributing much of the engagement to stories because viewers actively choose to watch them rather than being algorithmically fed the content.
- Kevin explains that Make.com cannot support multiple entry points in a single workflow, requiring the two-part automation (generation and publishing) to be split into two separate blueprint files, whereas n8n supports multiple triggers within the same workflow.
- Kevin argues that using a higher-quality image model like Nano Banana (Imagen 3) significantly reduces the likelihood of typos and text errors in AI-generated story images, directly addressing a failure shown live during the demo.
- Kevin describes the advanced version's AI agent prompt as instructing the model to generate three variations of the image prompt with 'slightly different style, mood or composition,' adding creative depth compared to the simple one-to-one prompt-to-image approach.
- Kevin recommends that users start by identifying what they are selling and what action they want viewers to take—such as visiting a site or sending a keyword—as the foundation for writing effective story prompts, arguing that straightforward automations with clear goals tend to be the most effective.
Topics
Full transcript available for MurmurCast members
Sign Up to Access