Make Money on Instagram WITHOUT Posting (No Code)

Alston built an Instagram account that posts itself. Every 12 hours, an AI picks a keyword from a spreadsheet, generates a scroll-stopping image, writes a caption, and pushes everything live on Instagram while he does nothing. The account was just 18 hours old when he ran the demo, and it was already working without any manual input.

This post walks through the entire N8N workflow he built, step by step. You will see exactly which tools he used, where the technical traps are, and how he plans to turn the growing audience into real income. No coding required. Just patience, the right sequence, and a willingness to push through one genuinely frustrating Facebook setup process.

What You’ll Walk Out With

A clear picture of how a fully automated Instagram posting system works from keyword to published post
The exact tools in Alston’s stack: Airtable, N8N, ChatGPT gpt-image-1, Amazon S3, and the Facebook Graph API
A step-by-step walkthrough of the N8N workflow including the nodes, the prompts, and the order they run
The specific gotchas that cost Alston five or six hours of debugging so you can skip them
Three monetization paths you can attach to this system once it’s running: affiliate links, PLR products, and digital courses
An honest look at what this costs, what it cannot do for free, and where you can potentially trim expenses
A direct path to finding what you should actually be building for your skills and situation at finder.platformproof.com

The Concept: What This System Actually Does

The idea is simpler than it sounds. You pick a niche, fill a spreadsheet with questions people search on Google, and let the automation handle everything else. Alston used the “can dogs eat” niche to test it. He loaded Airtable with keywords like “can dogs eat garbanzo beans,” set the workflow to run every 12 hours and 1 minute (the slight offset makes it look less robotic to Instagram), and watched the posts go live without touching anything.

The posts use AI-generated images that show a cute animal surrounded by the food from the keyword, with the question printed directly on the image as text. The image stops the scroll. The caption answers the question and drives engagement. The whole sequence runs on autopilot once it’s wired up.

Alston chose “can dogs eat” and “can cats eat” style niches because they generate genuinely useful, answerable content, attract a passionate pet-owner audience, and have real affiliate products attached. It is not a random niche choice. The audience will click on dog food recommendations and healthy treat guides, which is exactly what affiliate links need to convert.

Before You Build: What You Need First

Three things need to be in place before you touch N8N. Getting any one of them wrong will stop the workflow cold, and figuring out which one is the problem takes time you don’t want to waste.

An Instagram Professional account. Not creator. Not personal. Professional. The Facebook Graph API does not work with the other two account types. If you try to connect with a creator or personal account, nothing will post and you will spend hours wondering why. Go into your Instagram settings and switch to Professional before you start building anything.

A Facebook page linked to that Instagram account. This was the single biggest trap Alston ran into. He built the entire workflow for a test account called “doggy snacks,” launched it, and it failed completely. The reason: no Facebook business page was connected. Instagram business accounts need to be tied to a Facebook page through business.facebook.com, or the Graph API will not grant the permissions you need. Connect the page before you build anything else. This one step alone will save you the kind of frustration Alston described as a “huge pain.”

An account at developers.facebook.com. The Facebook developer portal is where you create your app, generate your access token, and add the Instagram permissions the workflow needs. Alston described this step as “really annoying,” but once you do it once you never have to redo it. The skills carry over to any future account or project you set up.

Setting Up Your Airtable Keyword Database

Airtable is the brain of the system. It stores every keyword the automation will work through, tracks what has already been posted, and saves the captions for later repurposing on other platforms.

Alston’s Airtable base had these specific columns: keyword (the full search query, like “can dogs eat garbanzo beans”), seed keyword (just the subject, like “garbanzo beans”), uploaded to IG (yes or no, defaulting to no), IG caption (saved for repurposing), a link field, date posted, raw image URL, and last modified date.

The “uploaded to IG” column is what the N8N trigger uses to decide which row to process next. It searches for any row where uploaded is set to no, grabs one record, runs the full workflow on it, then updates that field to yes when the post goes live. That loop is what keeps the system moving through your keyword list automatically without repeating content.

The seed keyword column deserves a specific note. Alston uses it to avoid spaces in filenames when uploading to Amazon S3. Full keyword phrases like “can dogs eat garbanzo beans” have spaces that will break the S3 URL if you’re not careful. The seed keyword “garbanzo beans” is shorter and easier to clean up. He also uses it in the image generation prompt to tell the AI what to surround the animal with, so the seed keyword does double duty.

Finding Keywords That Drive Traffic

You need a list of keywords before any of this runs. Alston used Ahrefs, a paid keyword research tool that also has a free tier. He went to the keyword generator, typed in “cans eat” as a root phrase, and got 20 to 30 keyword suggestions to get started. Not a massive list, but enough to prove the system works and start collecting data on what resonates.

His process for extracting seed keywords from that list: copy everything into Google Sheets, use Data > Split Text to Columns with “eat” as the separator, and the word after “eat” becomes your seed keyword. “Can dogs eat garbanzo beans” splits into “can dogs” on one side and “garbanzo beans” on the other. Copy the right column into Airtable as your seed keyword field.

Alston also mentioned using ChatGPT to generate keywords dynamically. You can ask any LLM for “a keyword for the niche can dogs eat, only return the keyword, nothing else” and get a fresh keyword with each workflow run. He started with a static list to make sure things worked first, but the dynamic option lets the system expand itself indefinitely without you adding rows manually.

The most important thing about keyword choice is the intent behind the search. Questions like “can dogs eat garbanzo beans” are informational queries with a specific audience. Pet owners searching this are actively researching what is safe for their animals. That audience will engage with content that genuinely answers the question, and they will click affiliate links for treats, food guides, and dog care products. That behavior is what makes the monetization work.

Generating AI Images with ChatGPT gpt-image-1

This is where the visual element comes from, and the model choice matters. Alston specifically uses ChatGPT’s gpt-image-1, not DALL-E 3. The reason is direct: DALL-E 3 cannot reliably put legible text inside an image. gpt-image-1 can. Since the whole concept depends on the keyword question appearing as readable text on the image, you need gpt-image-1. It is more expensive and slower than older models, but there is no shortcut around this requirement.

The prompt structure Alston uses is simple. For the garbanzo beans example, it was: “Create an image of a cute puppy surrounded by garbanzo beans. Include text on the screen that says can dogs eat garbanzo beans.” The output looks approximately 75% real, by his own description. Someone would have to look carefully to realize it is AI-generated. The image quality is high enough to stop the scroll, which is the primary job.

Set the resolution to 1024×1024. Alston also recommends checking “respond with URL” rather than letting the API return the raw base64 string directly, which simplifies the next step in the pipeline. He tested adding a watermark instruction to the prompt: “Please create and add a watermark that says owned by [account name].” The watermark appeared on the generated image. This matters once your content starts spreading. People will download and re-upload your images without credit. A visible watermark makes it harder to erase where the content came from and quietly builds your brand across every share.

One critical N8N habit from the walkthrough: pin your data during testing. Every time you test the image generation node without pinning, you spend real API tokens creating a new image. Pin the output of that node while you configure everything downstream. Unpin only when you are ready to do a complete end-to-end run. This single habit can save you meaningful money during the build phase, especially when you hit errors and need to retest repeatedly.

Converting Base64 and Uploading to Amazon S3

ChatGPT returns the image as a base64-encoded string. You cannot put that string directly into Instagram. You have to convert it to an actual image file first, then host it somewhere that returns a public URL, then pass that URL to the Facebook Graph API.

N8N has a built-in node that converts base64 to binary data. That binary becomes your image file. From there, Alston uploads it to Amazon S3. He creates an S3 bucket with public access enabled, names the file using the seed keyword to avoid spaces, and uploads it. The result is a public URL he can pass directly to Instagram’s API.

One S3 setup detail you need to know: when you create the bucket, AWS puts up multiple warnings about making contents publicly accessible. You have to click through them intentionally. The bucket access model needs to be “ACLs enabled” with the public access blocks removed. AWS tries to talk you out of this because public buckets can be a security risk in other contexts. For this use case, it is the correct setting. General purpose bucket, ACLs enabled, all objects in the bucket can be publicly accessible.

After the post goes live on Instagram, Alston deletes the file from S3 immediately. Amazon charges for storage, and he has no reason to keep the image in the bucket once it has been used. He saves the image to Google Drive before deletion if he wants a local copy, but the S3 file goes away right after posting to keep costs as low as possible.

Writing AI Captions That Build an Audience

The caption is generated by a second ChatGPT call, separate from the image generation. Alston’s caption prompt framework includes several specific instructions that matter for performance: keep it under 2,200 characters (Instagram’s character limit), use small and medium-sized hashtags rather than massive ones that are impossible to compete in, answer the question in a helpful tone, write at a fifth-grade reading level, avoid asterisks, and speak like a human. He also recommends telling the AI to “act as an expert veterinarian” or whatever the relevant professional is for your niche. That framing produces answers with more practical authority and specific detail.

The SEO angle is worth paying attention to. Instagram images with keyword-rich captions that genuinely answer a question can appear in Google image search results. A post about “can dogs eat garbanzo beans” has a real shot at showing up on Google for that search. Each Instagram post becomes a search asset, not just a social post. That dual exposure is part of why informational niches work so well for this model.

Alston saves each IG caption back into Airtable in the dedicated caption column. That column becomes a secondary content asset. Every caption is a piece of writing you can repurpose into a TikTok script, a Twitter thread, a YouTube short hook, or a Pinterest pin description without generating anything new. The system produces reusable content as a byproduct of its main job.

Connecting Instagram Through the Facebook Graph API

This is the part Alston described as the main frustration point. He spent five to six hours on it the day before filming the walkthrough. The technical steps are not complicated in isolation, but Facebook’s setup flow has several places where things fail silently or return confusing errors, and tracking down the root cause requires working through multiple possible explanations.

Here is the sequence. Go to developers.facebook.com and create an app. Go to the Graph API Explorer. Generate an access token and add the Instagram permissions. Alston listed several that are required, all of which start with “Instagram.” The default token you generate lasts only a couple of hours. You need to click “Extend Access Token” to convert it into a long-lived or lifetime token. That lifetime token is what you put into N8N so the workflow keeps running indefinitely without you having to refresh credentials.

The N8N workflow uses two Facebook Graph API nodes for posting. The first one builds the post: it sends the S3 image URL and the caption to Facebook as a media container. This does not publish anything yet. It creates a staged version. The second node then publishes that staged post. That two-step process is how the Instagram API works at the Graph API level. You prepare first, then publish. Alston’s workflow uses version 22 of the Graph API. Both steps use POST requests, and N8N has pre-built nodes for both.

The Instagram business account ID you get from the Graph API Explorer maps directly to the Instagram node in N8N. That ID is what tells N8N which account to post to. Make sure you are pulling the ID from the correct connected account. Alston hit an error at one point because he was using the ID from a different meta business page that was not connected to his Instagram account. Once he matched the right page to the right ID, it worked.

Not sure which online income model fits your actual situation?

Answer seven questions and get a specific recommendation at finder.platformproof.com.

How to Monetize the Account Once It Is Running

Alston outlined three concrete paths. None of them require a large audience before you see any return.

Affiliate marketing. This is the fastest starting point. Join Amazon’s affiliate program, find products related to your niche (dog treats, pet food guides, grooming tools), grab your affiliate links, and put them in your bio with a call to action like “click below for my number one recommended dog treat.” You do not need thousands of followers to make the first sale. You need people who are already looking for a recommendation and land on your profile after finding a useful post. Informational niches produce that kind of visitor naturally.

PLR products. PLR stands for private label rights. These are done-for-you ebooks, guides, and courses that you can legally buy, rebrand, and resell. Alston found one called “23 Natural Dog Food” on a PLR marketplace for $4.99. He could sell it for $7 as a standalone product. He could bundle several PLR guides together and sell the bundle for $17. Or he could give it away free as a lead magnet in exchange for an email address, then build a list to market to over time. All three options work with a small Instagram following.

Digital courses. ChatGPT can help you outline an entire course on a topic like “how to cook healthy snacks for your dog.” From there, you could record the narration yourself, use AI-generated voiceover, create text-on-screen slides, or turn the whole thing into a PowerPoint-style presentation. The production options are wide open and none of them require a studio or expensive equipment. Your Instagram audience becomes your built-in customer base for the course launch once it’s ready.

Alston’s immediate plan was to start with affiliate marketing while the account builds. Because the posting runs automatically, he can check in every few days to monitor growth and swap in better-converting links as he learns what the audience responds to. The automation frees up his attention for the strategic work rather than the daily production work.

Step-by-Step Build Checklist

Create an Instagram Professional account (not creator, not personal)
Create a Facebook page and link it to your Instagram account via business.facebook.com
Create a developer app at developers.facebook.com and add all required Instagram permissions
Generate a lifetime access token using the Extend Access Token option in the Graph API Explorer
Build your Airtable base with columns: keyword, seed keyword, uploaded to IG (default: no), IG caption
Populate 20 to 30 keywords using the Ahrefs keyword generator or ChatGPT dynamic generation
Create an Amazon S3 account and a bucket with ACLs enabled and public access blocks removed
Build the N8N workflow in this order: Schedule trigger (12h 1m) > Airtable search (uploaded = no) > gpt-image-1 image generation > base64 to binary conversion > S3 upload > ChatGPT caption generation > Facebook Graph API media container (build) > Facebook Graph API publish post > Airtable record update (uploaded = yes) > S3 file deletion
Pin each node’s output during testing to avoid re-running paid API calls while debugging
Set the schedule to active and verify the first automated post lands on the Instagram profile correctly

Honest Drawbacks

This system is not free. The OpenAI API for gpt-image-1 costs real money per image. Alston is direct about this: you cannot do this without spending on API tokens because image generation requires paid access. Caption generation also uses tokens. How much depends on your posting frequency, but if you’re running twice a day you should calculate the cost before committing to a niche.

Amazon S3 also has ongoing costs, though they are small. Alston’s practice of deleting each file immediately after posting keeps S3 costs near zero, but you still need an active AWS account with billing set up.

The Facebook setup is genuinely difficult the first time. Alston spent five or six hours on it. There are external tutorials that walk through it in 20 to 30 minutes each, and he links to one in his video description. You will likely hit at least one confusing error where the API returns nothing or something unexpected. The most common causes are a missing permission, a wrong API version, or an account that is not properly linked. Check those three things first before going deeper.

The AI images look about 75% real, by Alston’s own assessment. Some viewers will notice they are AI-generated. That is not a dealbreaker in informational niches where the value is the answer, not the photography. But if you move this system to a niche where authenticity matters to the audience, you may see lower engagement than expected.

Building an audience that converts to income takes time. This system automates content production, not trust. You still need to respond to comments, optimize your bio, and add genuine value beyond the automated posts. The automation gives you more time to do that work rather than eliminating the need for it.

Find Your X

Automated Instagram posting is one income path. It fits people who are comfortable with technical setup, have patience for the Facebook API process, and want a system that runs without daily attention once built. But it is not the right fit for everyone, and starting the wrong model wastes months of time and money.

Answer seven questions at finder.platformproof.com and get a specific recommendation matched to your actual situation, skills, and timeline. No guessing. No trying ten things before finding what works.

Frequently Asked Questions

Do I need to know how to code to build this system?

No. N8N is a no-code automation tool. Every step is built by connecting visual nodes and filling in configuration fields. The only technically demanding moments are generating your Facebook access token and setting S3 bucket permissions, both of which have detailed tutorials already available. Alston built the entire system without writing a single line of code.

Can I use a free image generation model instead of gpt-image-1?

You can use DALL-E 3 or other models for plain images, but if you want readable text printed inside the image, gpt-image-1 is the reliable option right now. Models that do not support in-image text will produce images without the keyword question visible, which removes one of the main reasons the posts perform well. If you find another model that handles in-image text reliably, it could substitute, but test it thoroughly before switching the full workflow over.

Why does my N8N workflow fail to post to Instagram?

The three most common reasons: your Instagram account is not set to Professional, your Facebook page is not connected to your Instagram account via business.facebook.com, or your access token has expired and needs to be extended. Check all three before going deeper into debugging. The Facebook page connection is the one most people miss because nothing in the N8N error message tells you that is the problem.

How much does this cost per month to run?

The main costs are OpenAI API calls for gpt-image-1 image generation and caption generation, plus a minimal amount for Amazon S3. Posting twice a day means roughly 60 images per month. gpt-image-1 pricing varies by resolution and output type. At standard rates this is not a zero-cost system. Run it at low frequency for the first two weeks, check your actual API usage in the OpenAI dashboard, and then decide whether to increase the posting schedule based on real numbers.

What niches work best for this type of account?

Informational question niches work best. “Can dogs eat X,” “can cats eat X,” “is X safe for babies,” “how to fix X at home” all follow the same pattern: high search volume, answerable questions, and audiences that buy products related to the topic. Avoid niches where the audience expects personal stories, original photography, or behind-the-scenes content, since the AI-generated images will feel out of place in those contexts.

How long before I see my first affiliate commission?

There is no fixed timeline. Alston’s account was 18 hours old when he filmed the walkthrough. First commissions depend on how quickly posts gain search visibility, how many people click your bio link, and whether the affiliate products match what the audience is already looking for. Starting with Amazon affiliate links in a product-heavy niche gives you the fastest feedback loop because Amazon has a massive product catalog and a program almost anyone can join the same day.

What is a PLR product and where do I find them?

PLR stands for private label rights. These are digital products (ebooks, guides, templates, mini-courses) that you can legally buy, edit, and resell under your own name. PLR marketplaces sell them for a few dollars to around $30. Alston found a “23 Natural Dog Food” guide for $4.99. You can resell it for $7 as-is, bundle several guides and sell the package for $17, or give it away free to collect email addresses and build a list. Search “PLR” plus your niche keyword to find options in your category.

Can I run this system for multiple niches at the same time?

Yes, but each niche needs its own Instagram Professional account, its own Facebook page connected to that account, and its own Airtable base. You can duplicate the N8N workflow and update the credentials and Airtable connections for each new account. The setup cost is front-loaded. Once one account is working, the second is faster to build because you already know the process and have already survived the Facebook API configuration once.

Sources

Alston Godbolt, “Make Money on Instagram WITHOUT Posting (No Code)”: https://youtu.be/_uvMovE-4o8
N8N workflow automation platform: n8n.io
Airtable database platform: airtable.com
OpenAI gpt-image-1 image generation API: platform.openai.com
Amazon S3 cloud storage: aws.amazon.com/s3
Facebook Graph API and developer portal: developers.facebook.com
Ahrefs keyword generator tool: ahrefs.com