Can ChatGPT Analyze an Instagram Profile? (2026)

Can ChatGPT analyze an Instagram profile?#

On its own, no: ChatGPT can't open Instagram profiles. Ask it about @someaccount and it will either say it has no access or guess from whatever it remembers from training, which is usually old and often wrong. Connect StoriesFly to it and it reads live public data (the profile, posts, reels, stories, highlights, comments and follower lists) and analyzes that. The same works in Claude, Claude Code and Cursor.

Updated October 2026.

This guide covers why ChatGPT can't see Instagram by default, how to connect it in about a minute, a worked example with the exact prompts and what they cost, and what no AI tool can tell you about an account from the outside.

Why ChatGPT can't see Instagram by default#

There are three reasons, and they stack.

1. Instagram shows little to visitors who are not signed in. Follower lists, comments and older posts sit behind a login prompt. ChatGPT's browsing does not sign in to Instagram, so at best it sees a fragment of the page.
2. Training data is old. Even when ChatGPT "knows" an account, what it knows is a snapshot from months or years ago. Follower counts, posting rhythm and recent content change every week.
3. Screenshots are slow and partial. You can paste screenshots, but then you are the one collecting the data. Twelve screenshots of a grid still leave out view counts, comment text and dates.

What fixes this is a connector. ChatGPT and Claude both support MCP (Model Context Protocol), the standard way to plug a tool into an assistant. Once a connector is added, the assistant can call it in the middle of a chat, get structured data back and reason over it.

How to connect Instagram to ChatGPT#

The full steps with screenshots are on the connect page. In short:

ChatGPT (works on the Free plan too):

1. Open Plugins, then Add, then Add custom MCP server.
2. Paste the address https://storiesfly.com/mcp and pick OAuth for sign-in.
3. Press Connect and sign in with Telegram on the StoriesFly page.

Claude (claude.ai and the apps):

1. Open Customize, then Connectors, then Add custom connector.
2. Name it StoriesFly and paste the same address.
3. Press Connect and sign in with Telegram.

Claude Code and Cursor take the same address; the commands are on the connect page. There is no API key to copy in any of them.

Your first 100 requests are free, once per Telegram account. That is enough for about ten full profile analyses. You never log in to Instagram with your own account, and your account never appears among a story's viewers.

A worked example: analyzing a public profile#

Here is a real flow, with a made-up account name. Start a new chat after connecting and ask:

Analyze the Instagram account @northtrail.coffee. What do they post, how often, what works best, and what are people saying in the comments?

The assistant decides which tools to call. For a question like this it usually makes these calls:

What comes back is a short report. A typical answer covers:

• Posting rhythm. "4.1 posts a week over the last 30 posts, with a two-week gap in August."
• Format mix. "60% reels, 30% carousels, 10% single photos. Reels get a median of 3× the views of photos."
• What works. The top five posts by engagement, with what they have in common ("behind-the-bar videos under 20 seconds").
• Engagement rate. Likes plus comments divided by followers, per post and as a median, so one viral post does not skew it.
• Comment themes. Questions about prices, requests for a location, emoji-only replies.
• What to try. A few concrete ideas drawn from the above.

You can then follow up without starting again:

Compare that with @harbor.roasters. Same metrics, side by side in a table.

That is roughly another 10 requests. Or narrow it:

Only reels from the last 60 days. Which hooks in the first line of the caption get the most views?

Prompts that work well#

Be specific about the window and the output. Assistants read more when the question is vague, and every page is a request.

Quick profile check (about 2–3 requests):

Look up @account on Instagram. Give me follower count, posts per week
over the last 9 posts, and median likes per post. Keep it to 5 lines.

Full audit (about 10 requests):

Do a content audit of @account: posting rhythm, format mix, top 5 posts
by engagement rate and why they worked, comment themes on the best post,
and 3 things to test next month. Use the last 30 posts.

Your own account (about 10 requests):

This is my account, @account. Find the posts that underperformed my own
median by more than 50% and tell me what they have in common.

Hashtag check (about 3 requests):

How many posts use #specialtycoffee, and what do the top posts under it
look like right now? Read one page of top posts.

Follower sample (about 4 requests):

Read the first 100 followers of @account and tell me roughly how many look
like real people vs. empty or brand-new accounts.

A followers page holds up to 25 accounts and costs 1 request, so 100 followers is 4 requests.

What the assistant can and cannot see#

It sees what a public profile shows: the profile, posts and reels with their like, comment and view counts, captions and dates, stories and highlights, comments, a sample of likers, follower and following lists, hashtags and search. For a single post it can read shares, saves and reposts where Instagram gives them.

It cannot see:

• Private accounts. Nothing beyond the profile header. The assistant will tell you the account is private; that look-up counts as one request.
• Reach and impressions. Only the account owner sees these in Insights.
• Audience demographics. Age, gender and location of followers are in Insights only.
• Profile visits, link clicks, saves on every post. Again, owner-only data.
• Full follower lists of big verified accounts. Instagram shows only about 50 followers of a verified account.
• Who viewed a story. Only the poster sees that.

So treat any AI "analysis" of someone else's account as an analysis of public signals. It is good at rhythm, formats, content patterns and engagement; it is guessing on reach and audience. If you need Insights for your own account, export them from Instagram and paste them in alongside.

How much it costs#

The free requests and plans are shared with the StoriesFly API. The balance never expires, and if you run out of a plan mid-month, the assistant can keep going on the balance. Details are on the MCP page.

When to use something else#

• You only need your own account's numbers. Instagram's own Insights, or a tool that connects through Meta's official login, will give you reach, demographics and saves that no outside tool can see.
• You need posting or scheduling. StoriesFly reads public data only. It does not post, like, follow or send messages.
• You need the whole internet's mentions of a brand. That is a social listening tool, not a profile reader.

For a broader look at the options, see the comparison of MCP servers for Instagram and TikTok. For TikTok, the same connection works: see how to analyze a TikTok account with ChatGPT or Claude. For the manual method without AI, there is the Instagram competitor analysis guide.