It lives at igunfollowed.com and runs entirely in the visitor's browser.

The trick: Instagram already gives you the data

Instagram lets you download your own data (Settings → Accounts Centre → Your information and permissions → Download your information). Inside that export there's a followers_and_following folder with two files:

• followers_1.html (or JSON) — everyone who follows you
• following.html (or JSON) — everyone you follow

Anyone in list B who isn't in list A doesn't follow you back. That's the whole analysis. No scraping, no login, no API keys.

Doing it locally, in the browser

The manual way is painful: two long HTML files and a generic list-diff tool. So the tool automates exactly that, client-side:

1. The user drops the official export ZIP (or the two files) onto the page.
2. JavaScript unzips and parses them in the browser — the files never leave the device.
3. A set difference produces the non-followers, mutuals, and fans lists, with snapshot comparison and CSV export.

Because nothing is uploaded, the architecture is simple: a static site, zero backend, zero database, zero credentials to leak. Privacy isn't a policy page here — it's the fact that there's no server to send data to.

What I'd do differently

• The export format changes occasionally (HTML vs JSON, file naming), so the parser has to be forgiving.
• Explaining "request your data from Instagram" to non-technical users is the biggest UX hurdle — most people have never opened that settings screen.

If you try it, I'd genuinely like feedback on the parsing edge cases: igunfollowed.com