How to Connect Instagram Public Data to Gemini for Competitor Analysis

To analyze competitors’ Instagram accounts in Gemini, add their usernames to the Instagram Public connector in Windsor.ai, then add Windsor.ai as a data store in Gemini Enterprise and authorize it once. Gemini can then compare follower counts, posting frequency, formats, captions and likes and comments across the accounts you track.

The data is what any visitor can see on a public profile: profile counts, every public post with its caption, and the likes and comments on each post. You do not log in to the accounts you track, and they do not have to approve anything.

This guide is for accounts you do not own. For reach, saves and Reels watch time on your own account, see how to connect Instagram to Gemini.

Start tracking a public Instagram account on the free forever plan, no credit card: onboard.windsor.ai/app/instagram_public.

Which accounts you can track

Windsor.ai reads public profiles through Meta’s business discovery endpoint in the Instagram API, which sets three conditions.

• You need your own Instagram Business or Creator account. Windsor.ai runs each lookup from your own account, through the Facebook login that manages it.
• Tracked accounts must be Business or Creator accounts. Meta’s API returns data about other professional accounts. Private accounts and most personal accounts return nothing.
• Age-gated business accounts are not returned. Accounts that restrict their audience by age come back empty.

If a username returns no data in Gemini, check these three conditions first.

Before you start

• A Windsor.ai account. The free forever plan covers one tracked profile (one data source, one account) with 30 days of history. Tracking several competitors, or adding your own Instagram Insights and ad data next to them, needs a paid plan: from $19 a month (billed annually) for 3 sources and 75 accounts, or $99 a month for 7 sources. See Windsor.ai pricing.
• Gemini Enterprise Standard, Plus, Frontline or Pay-as-you-go, set up in the Google Cloud console. The Business edition has its own setup and fewer connectors.
• A Google Cloud project under your organization (Google Workspace or Cloud Identity, not a personal Gmail account), with billing on, Gemini Enterprise activated and the Discovery Engine Editor role for the person doing the setup.
• The usernames of the accounts you want to track.

Google bills Gemini Enterprise and the data store separately from Windsor.ai.

Step 1: Add the public profiles in Windsor.ai

1. Go to onboard.windsor.ai/app/instagram_public and log in or create an account.
2. Sign in with the Facebook login that manages your own Instagram Business or Creator account.
3. Type the usernames of the accounts you want to track and save them.

Each tracked username counts as one account on your Windsor.ai plan.

Step 2: Add Windsor.ai to Gemini Enterprise

Google calls this connection a data store. The main actions are below; the Gemini Enterprise setup guide has every screen and the common errors.

1. In the Google Cloud console, open Gemini Enterprise, click Data stores, then Create data store. Under Third-party data sources, find Windsor.ai and click Add data source.
2. Under Entities to search, select Connectors. On the Actions step you can leave both actions off: Instagram Public Data is read only (see the Windsor.ai write actions list).
3. Under Multi-region, keep global unless your company must keep data in the US or EU. This cannot be changed later. Name the data store and, if prompted, configure access control with Google Identity.
4. Keep General pricing, click Create, and wait for the state to change from Creating to Active.
5. In your Gemini Enterprise app (in the same location), open Connected data stores, click Add existing data stores, select the Windsor.ai data store and click Connect.
6. Open the app’s web app URL, click the puzzle piece icon below the Ask Gemini Enterprise box, click Authorize next to Windsor.ai, then Allow Access, sign in to Windsor.ai and click Finish. Make sure the Windsor.ai toggle is on.

The Windsor.ai data store is in public preview under Google’s Pre-GA terms. If this is the first Gemini Enterprise app in your project, Google adds a 30-day Gemini Enterprise licence with it.

Step 3: Check the data

Start with a profile snapshot you can compare against the Instagram app:

For each Instagram account I track, show profile_username, profile_followers_count, profile_media_count, biography and website in one table.

Follower, following and post counts are today’s values, with no history. Post-level likes and comments are available for every public post in your date range.

Instagram Public Data fields

Area	Fields	What you can do with them
Profile	profile_username, profile_name, profile_followers_count, profile_follows_count, profile_media_count, biography, website	Size and positioning comparison
Posts	media_timestamp, media_type, media_product_type, media_caption, media_permalink, media_url, media_thumbnail_url	Content audits, format and topic analysis
Engagement per post	media_like_count, media_comments_count	Top posts, engagement per post
Averages	likes_per_post, comments_per_post	Quick benchmark across accounts

There is no reach, views, saves, shares or audience data for accounts you do not own. Meta only reports those to the account itself.

Prompts for competitor analysis in Gemini

Posting frequency and format mix

For @brand_a, @brand_b and @brand_c, count posts per week over the last 30 days by media_type. Who posts most, and which format does each account rely on?

Top posts per competitor

List the 5 posts with the most likes plus comments for each account I track in the last 30 days, with caption, type, date and permalink. What do the top posts have in common?

Caption themes

Group @competitor’s captions from the last 30 days into topics. Which topics get the most comments per post?

Engagement against audience size

Divide average likes and comments per post by follower count for each account I track. Which account gets the most engagement for its size?

Creator check before a sponsorship

For @creator, show posts in the last 30 days, average likes and comments, how many posts are sponsored according to the caption, and whether recent posts do better or worse than older ones.

On the free plan, prompts can cover the last 30 days. Longer periods need a paid plan.

Compare competitors with your own data

With your own sources connected in Windsor.ai, Gemini can put the outside view next to your inside view:

• Your Instagram Insights: compare your posting frequency and engagement per post with each competitor, using the same fields.
• Meta Ads: check whether a competitor’s posting push lines up with a rise in your cost per result.
• TikTok Organic or YouTube: see whether the same creator or competitor performs differently on another platform you track.

Troubleshooting

Windsor.ai finds no Instagram account to run the lookup from

Meta documents business discovery as part of its Instagram API with Facebook Login, and that route needs your own Business or Creator account to be connected to a Facebook Page. Sign in with the Facebook profile that has access to that Page. An Instagram-only login, or a professional account with no Page attached, leaves Windsor.ai with no account to run the lookup from.

A post shows no likes, or fewer likes than the Instagram app

Meta omits the like count from lookups of another account’s posts when the owner has hidden like counts, and media_like_count never includes likes on promoted versions of a post (Meta’s Instagram media reference). For accounts that hide likes, compare media_comments_count instead.

Gemini returns nothing for the accounts you track

After a new data store is connected to a Gemini Enterprise app, results can take about 10 minutes to appear. If the answer is still empty after that, open the puzzle piece menu in the web app and turn the Windsor.ai toggle on, or click Authorize if it is still showing, and confirm the data store state is Active in the Google Cloud console.

Next steps

Add the usernames at onboard.windsor.ai/app/instagram_public, then set up the data store with the Gemini Enterprise setup guide. For more questions, see the Windsor.ai prompt library. The same competitor workflow in Claude is covered in Instagram competitor analysis with Claude, and other Gemini sources are listed on the Gemini Enterprise integrations page. For ChatGPT or Microsoft Copilot, see how to connect Instagram Public Data to ChatGPT and how to connect Instagram Public Data to Copilot.