Who can see your saved reels: a privacy walkthrough

Putting a tool between yourself and the things you save is a small act of trust, and it is one people tend to perform without examining. What you are handing over is not dramatic — a list of videos you found interesting — but a list like that is unusually revealing. It says what you are worried about, what you are shopping for, what you are trying to learn and what you were doing at three in the morning.

Any product in this position should be able to answer four questions without hedging: what is stored, who can see it, who owns it, and how do you get rid of it. This is those four answers for Reel Mind, written to be specific enough to check. Where the product does less than you might assume, that is said rather than skirted.

What gets stored

When you share a reel to the bot, the record that gets written is not the video. It is a link to the original Instagram post, the media ID, and the ID of the account that sent it — which is you. Processing then adds derived text to that same record: a transcript of the spoken audio, the on-screen text pulled out by OCR, the creator’s caption, a category with a confidence score, and a small set of representative frames chosen from the video.

The distinction that matters most is that the video itself is never re-hosted. A reel in your library is a pointer to a post on Instagram plus a description of what that post contained. Nothing plays from our servers, and there is no copy of the file sitting in storage waiting to be found. That is a privacy property and a limitation at the same time, and the same sentence covers both.

On the account itself, what exists is a username, a securely hashed password, and the Instagram user ID the account is bound to. That ID is the single thread joining the two halves of the product, and it is worth understanding as a privacy fact rather than a technical one: it means the system knows your Instagram identity, because it must, and knows nothing else about that account beyond it.

Separately from any of that, the website behaves like a website: there is a session cookie that keeps you signed in, and analytics and advertising cookies from Google. That is ordinary site-level data rather than anything about your library, and the privacy policy lists the providers involved and what each one receives.

That covers what is kept. The other half of the question is who handles a reel while it is being worked on. The transcription, OCR and categorization models run inside services we operate: reel content is not sent to an external AI provider’s API, and it is not used to train anything. Those services are hosted on Hugging Face, so Hugging Face handles a reel’s audio and its extracted frames on our behalf for as long as processing runs — the relationship any hosting provider has with what runs on its machines. One thing does cross that boundary. To pin the category down more accurately, a short excerpt of the reel’s text is sent to DuckDuckGo as a search query and the results are used as topic context; no account identifier travels with it.

What the public Explore page shows

There is a public page — Explore — that anyone can open without an account, showing a curated preview of saved reels grouped by category. It exists so people can see what a filed library looks like before creating one, and because it is public it is the part of the product most worth being precise about.

Each entry on that page is a link to the original Instagram post, its caption, and Instagram’s own thumbnail image. That is the whole payload. No username, no account ID, no transcript, no extracted on-screen text, no saved frames — nothing that connects a reel to the person whose library it came from. Somebody browsing Explore is looking at reels; they are not looking at anyone’s collection, and the page provides no route from a reel back to a saver.

You can also be left out of it entirely. Ask to be excluded and your reels stop appearing there while your dashboard carries on unchanged. Worth knowing about the timing: the page is cached, so an exclusion takes effect at the next scheduled rebuild rather than the instant the request is read.

Who owns the content

The creator does, permanently, and nothing about saving a reel changes that. Reel Mind claims no ownership of any reel. Copyright and every other right stay with whoever made the video, and passing a link through an organizing tool is not a transfer of anything — it could not be, and it is not being quietly attempted in a terms document.

This is the practical reason the product stores a pointer rather than a copy. Re-uploading someone’s video would put a second version into the world with the views, the attribution and the control detached from the person who made it. A link keeps all three where they belong: you click through to the original post, the creator gets the traffic, and if they change or remove the video their decision propagates automatically.

It also means a removal request from a creator is a legitimate request rather than an awkward one. If you made a reel that appears here and you want it gone, that is your call to make and no account is needed to make it. The terms say so and the contact page has a route for exactly that.

Deleting things

Removing a single reel is self-serve and immediate. Delete it from your dashboard and the whole record goes — the link, the transcript, the extracted on-screen text, the stored frames, all of it, because they live on one document rather than scattered across several. There is no archive it moves into and no tombstone left behind, so it is worth confirming you mean it when the dialog asks.

Closing an account is not self-serve today, and it would be misleading to imply otherwise. There is no delete-my-account button in the dashboard. What exists is a request route: you email a deletion request from the contact page, and the account and the reels saved against it are removed. The privacy policy also describes revoking the app’s access from your Instagram account settings, which sends a deletion request to our servers.

What is deliberately not being claimed here is a timescale. This article is not going to state a retention window or a turnaround guarantee, because a deletion request is handled by a person rather than by an automated job with a clock on it — and a promise the code does not enforce is worth less than an honest description of a manual process. If a hard deadline matters for your situation, say so in the request rather than assuming one applies.

None of the above is remarkable, and that is rather the point. The reason to write it down is that the alternative — a policy page nobody reads and a product whose behaviour has to be inferred — leaves you guessing about a system holding a fairly intimate list. Guessing usually resolves in one of two unhelpful directions: assuming a tool does more than it does, or abandoning something useful over a fear of something that was never happening. A description specific enough to be checked is what lets you skip both and decide on the actual facts.