How to Do an Instagram Audit With AI for Less Than €1

Most advice on how to do an Instagram audit tells you to open your own profile, scroll your last thirty posts, and write down what you notice. That is not an audit. That is a feeling. You cannot see your own median reach by scrolling, you cannot count your own caption length by eye, and you certainly cannot compare any of it against the people beating you.

I ran a real one last week. It pulled 70 posts across three accounts, worked out every number I could not see by hand, and cost less than the price of a coffee. Here is exactly how it works, including the part where it told me something about my own account I did not want to hear.

What an Instagram audit is actually for

An audit has one job: to replace what you assume with what is measurably true.

Almost everyone gets this wrong in the same direction. They assume the problem is volume. They think they are not posting enough, so they resolve to post more, and nothing changes, because volume was never the constraint.

The only way to know is to put your numbers beside somebody else's. Not a benchmark from a blog post written two years ago in a different niche. The actual accounts you are competing with, measured over the same window as you.

That means you need four things:

• Every post from your own account over a recent window, with plays, likes, comments and caption length
• The same data for two or three competitors in your exact niche
• The medians, not the averages, because one viral post will lie to you
• Somebody, or something, to read all of it and tell you what differs

The first three are a data problem. The fourth used to be the expensive part.

The tool that does the pulling

Instagram will not hand you a competitor's post history, so you need a scraper. I use the Apify Instagram Scraper, which is a maintained public actor rather than somebody's abandoned side project.

The thing to check before you pick any scraper is coverage. A lot of them only return standard posts, which is useless now. You need one that pulls reels, carousels, posts and profiles together, because the whole point of the exercise is to find out which format is doing the work.

Pricing is per result rather than per month, which is the reason this is affordable at all. You are not renting a seat on a platform. You are paying for the rows you actually pull. Apify also gives you five dollars of free credit when you sign up, which is more than enough to run this several times before you spend anything.

If you would rather compare hosted options first, I have written separately about AI Instagram audit tools and what they do and do not cover.

Step 1: Connect the scraper to your AI

The scraper returns raw JSON. Hundreds of rows of it. Reading that by hand is worse than scrolling your feed, so the second half of this setup is connecting it to an assistant that can do the arithmetic.

I use Claude, connected to Apify through an MCP connector. The process takes about two minutes:

1. In Apify, open Settings, then API and integrations
2. Copy your API token, or add the MCP connector directly
3. Back in Claude, run /mcp to confirm the connection is live

Once it shows as connected, Apify is a tool your assistant can call on its own. You stop copying files between tabs, which is the step where most people quietly abandon this kind of workflow.

Step 2: Write the prompt that does the work

This is the part that decides whether you get an audit or a pile of data.

Name three things explicitly. Name the accounts you want scraped, including your own. Name the scraper you want used, by its URL. Name the output you want back, in the shape you want it.

My prompt asked for a competitor analysis across two accounts in my niche plus my own for benchmarking, using the Apify connector and that specific scraper, returned as a live dashboard with a summary at the top and a ranked list of fixes at the bottom.

Including your own account is not optional. A competitor report with no baseline tells you what other people do. It does not tell you what to change.

Be specific about what you want to see. A ranked list of fixes is more useful than a wall of charts, because the ranking forces a decision about what to do first.

Step 3: Read the numbers that matter

The run came back with 70 posts across three accounts: mine at 269 followers, one at 38,000, and one at roughly 294,000.

Here is what it found, and it was not what I expected.

I was not posting too little. Over the window measured I shipped 11.7 posts a week. The account with 294,000 followers shipped 10.5. I was out-working someone with a thousand times my audience and getting nowhere, which rules out effort as the explanation and forces you to look at everything else.

The gaps that mattered were all in how the posts were built:

Measure	Me	The larger accounts
Median caption length	501 characters	39 characters
Hashtags per post	Four to five	Zero, on all 48 of their posts
Median comments	Zero	Hundreds
Carousels	1 of 22 posts	13 of 24, on one account

Zero hashtags across every single competitor post is the kind of finding you would never reach by scrolling. It is not a trend piece or an opinion. It is 48 out of 48.

Median caption length is the gap that surprised me most. Mine ran to 501 characters, which is a short blog post sitting under a video. Theirs ran to 39, which is one sentence. I had assumed a longer caption was more generous, and that teaching in the caption was worth doing. What the data says is that nobody scrolling stops to read it, so the effort lands on an audience of nobody.

The comments column is the one that actually explains the reach gap. Instagram treats comments as a stronger signal than likes, and one of these accounts was earning more comments than likes because every caption ended by asking for a specific word in the replies. My captions were teaching. Theirs were asking.

Step 4: Look for the proof already in your own account

The most useful thing an audit does is find the experiment you already ran and forgot about.

Mine was sitting there in August. Two reels, ten and eighteen seconds long, with a one line caption and a comment keyword. They did 1,666 and 1,267 plays against a median of 109. Every reel after that went back to twenty six to sixty two seconds with a paragraph caption and four hashtags, and reach fell back to double digits.

I had found the format and then stopped using it. Without pulling the data I would not have noticed, because two good posts in a month feel like luck rather than a result.

Look for this pattern in your own numbers. An outlier that matches a change in method is not luck. It is a repeatable thing you abandoned.

Step 5: Turn it into a short list you will actually do

An audit that produces twenty recommendations produces nothing. Ask for a ranked list and keep it to five.

Mine came back as: cut the caption to one line with a keyword in it, delete the hashtags, get reels under twenty seconds, ship three carousels a week, and lock the posting time.

Every one of those is tied to a measured gap rather than to general advice. That is the difference between an audit and a checklist you found online. If a recommendation cannot point at a number in your own data, it does not belong on the list.

What it costs

This is the part that surprises people.

The run I did the day before cost 38 cent. The one in the video moved my usage from about 33 cent to 57 cent, so roughly 24 cent for 70 posts across three accounts. Apify bills in dollars, so in euro terms you are still comfortably under a euro either way.

Compare that to a competitor research platform, where the entry tier usually starts somewhere around a hundred a month. For an audit you run once a quarter, a subscription is the wrong shape entirely. You want to pay for the pull and then stop paying.

The five dollars of free credit covers your first several runs, so you can test the whole thing before you spend anything at all.

Where this breaks

It would be dishonest to leave this here, so here is the other side.

It will not capture everything. My trial reels never appeared in the dataset. If you run experimental formats, check whether they are showing up before you draw conclusions from what is missing.

You still need to check the figures. The dashboard is generated from scraped data, and scraped data has gaps. Treat it as a strong first draft rather than an accounting record. Where a number looks wrong, go and look at the post.

Date filters are unreliable. Asking for posts newer than a given date is best effort, and older posts come back anyway. If you are running this repeatedly, you need your own record of what you have already seen.

Three competitors is thin. Two comparison accounts told me a lot, but four or five would tell me more, and would stop one unusual account skewing the whole picture.

The output is only as good as the accounts you choose. Pick competitors who are genuinely ahead of you in the same niche and format. An account that simply makes you feel behind, in a different business with a different audience, will produce a tidy report full of changes that do not apply to you.

Nobody is going to run it for you. I built this in July, ran it once, and never put it on a schedule. That is the honest reason to keep it cheap and manual rather than pretending it is an automation.

The same method works elsewhere

Once the connector is in place, the scraper is the only variable. The same setup covers LinkedIn, TikTok and YouTube, and there is a separate actor for YouTube comments that is worth a look on its own.

Scraping the comments under big videos in your niche is one of the better content research methods going, because a question asked repeatedly in a comment section is a video nobody has made properly yet.

If growing the account is the actual goal behind the audit, the mechanics of that are covered in more detail in my guide to growing Instagram followers organically.

Start with the question, not the tool

The reason most audits change nothing is that they produce description instead of decisions. You end up with a document that says your engagement rate is 1.2% and no idea what to do on Monday.

Ask a question you can act on. Why does this account get more comments than me on fewer posts? What do they do in the caption that I do not? Which of my own posts already beat my median, and what was different about them?

Then pull enough data to answer it properly, which now costs less than a euro.

If you would rather have someone run this on your accounts and hand you the shortlist, that is the kind of thing I do in a marketing consultancy session.

---

Disclosure: Some links in this article are referral links. If you use one, I may earn a commission at no extra cost to you. I only link to tools I actually use.

How this was made: The facts, numbers and examples in this article come from my own work and videos. I use AI to help draft and structure the writing, then I review and edit every line myself before it goes live.