Skip to content
Effective 2026-11-03

Privacy Policy

Video2Any has two data modes, plus an optional copy of finished decks in a signed-in account. This page states them separately, then lists accounts, payments, usage records, retention and who else can see data. Claims match what the product does today.

The short version

  • Mode 1 — Slide extraction. When you convert a video to slides, the video stays on your device. There is no video upload endpoint for extraction.
  • Mode 2 — Spoken notes and transcript. Only when you start notes or a transcript do we receive the audio soundtrack (and, for notes with pages, still images of those pages). We send the audio to third-party AI service providers to transcribe it. For notes (not transcript-only), we send the transcript and those page images to third-party AI service providers to write the handout. Finished notes and transcripts stay on your account until you delete them.
  • Library sync (signed-in). When you are signed in and the sync switch is on — it is on by default — the finished slide deck (the page images and their timing, never the source video) is stored in your account so it opens on another device. Free accounts keep the latest three decks. Paid accounts share a 100 MB allowance with saved notes, and older decks are removed first when it is full. The switch is in Account. Turning it off stops new uploads; deleting a deck or the account deletes the stored copy.
  • We also receive usage counts, page views, clicks, error reports, and — if you sign in or pay — account and purchase records.

Mode 1 — Slide extraction (runs on your device)

What is processed. Uploaded files, direct file links, screen recordings, and everything the converter produces for slides (pages, decks, frames) are processed inside your browser tab.

What is sent. For slide extraction there is no video upload endpoint. The video file is never uploaded. To keep work available when you come back, the video and the slides stay in your browser's storage on this device. Clearing the site's data removes them. A finished deck is also stored in your account when you are signed in with library sync on; that copy contains the slide images and their timing, not the video. See Library sync below.

What we still receive in Mode 1. Usage counts only: length, format, resolution, how long extraction took, how many pages came out, what you changed, which format you exported, which pages you opened, which buttons you clicked, and error reports. These use a random browser id, and your account id if you are signed in. They do not include the video or the page images.

The browser extension that helps capture on-screen video follows the same Mode 1 rule for slide capture: it reads still frames in the tab and builds the deck on your machine. Starting Mode 2 from that deck is a separate action and follows Mode 2 below.

Mode 2 — Spoken notes and transcript (upload and AI processing)

Mode 2 starts only when you ask for a transcript or for notes. Nothing in Mode 2 is uploaded until you begin that job.

What leaves your device.

  • The audio soundtrack used for the job (taken from a video or audio file on your device). The full video file is not uploaded.
  • When the job is notes and the recording has pages, still images of those pages.
  • An optional title you set for the job.

Where it sits on our side. Audio and page images are stored in our private object storage under a per-job working prefix (notes/<account>/<job>/ on Cloudflare R2). That storage is not public.

Who processes it.

  • Transcription. Our servers read the stored audio, split it into overlapping chunks in memory, and send those chunks to third-party AI service providers to produce the transcript. This applies to transcript-only jobs and to notes jobs.
  • Notes writing. For notes jobs only: the transcript text and, when the job has pages, the page images are sent to third-party AI service providers to write the handout. A transcript-only job does not send page images.

How long working files stay. When a Mode 2 job succeeds, we delete its entire working prefix (notes/<account>/<job>/): the uploaded audio, page images, aligned transcript and every other working file. If the job fails, we keep the prefix so Retry can run without a new upload. A job that stops updating for about two hours is marked failed and its audio and diagnostic trace are deleted. Anything left under notes/is also configured to expire automatically within about two days. The finished text is kept separately under notes-text/ (see below), including the transcript of a transcript-only job. The working copy of the page images is deleted with that prefix; a copy of the pages the saved handout actually shows is kept under notes-frames/so an unlisted share link can still display them.

What stays on your account. The finished notes and/or transcript are stored under notes-text/. The page images the handout cites are kept undernotes-frames/. Both count toward the same storage allowance shown in the app (about 100 MB) and neither has an automatic time expiry. You can delete any saved result in the app; that deletes the saved text, the kept images, its unlisted share links, and that job's diagnostic trace. You can also create an unlisted share link so anyone with the link can read the handout, with its pages, without an account; revoking the link ends that access, and deleting the result removes the text and images it pointed at.

Each job may keep a diagnostic trace under traces/ (including transcript and notes text passed between the model and its tools). Deleting the saved result removes that trace, and the bucket is configured to expire traces within about thirty days.

What can leave, and only if you ask for it

None of these runs on its own. Each is a button you press, on one deck at a time.

Deck share links (retired)

Deck share links were retired on 2026-10-05, and new ones can no longer be made. Links made before then keep working until their expiry date, at most 30 days after they were made, and you can delete one sooner from your library. Expiry and deletion remove the stored deck, not just the link. A public link's deck and file name were uploaded as they were, so we could read them; a password link's deck was encrypted in your browser with a password we never received. For each link we keep its size, expiry date, any limit on opens, and how many times it was opened and a file saved from it, until the link itself is gone.

Saved decks (retired)

The library, where a signed-in user could keep a deck on our servers, was retired on 2026-10-05, and nothing new can be saved to it. Decks saved before then are stored unencrypted on our servers, with their title, slide count and video length, so they could be opened on another machine; each holds the finished slides and never the source video. They stay readable until 2026-11-04. Their owners were emailed a copy as PDFs; a copy too large to attach was placed behind a private link that works until 2026-11-04. On 2026-11-04 every saved deck, its search index, its folders and those copies are deleted.

Library sync (signed-in)

A new results library lets a signed-in account keep finished decks and open them on another device. It stores the deck archive — the slide images, their timing and the title — under library/<account>/<deck>/ on Cloudflare R2, with the title, slide count, video length and size in our database. The source video, PPTX and PDF files are not uploaded; exports are still made in your browser.

Sync is on by default for signed-in accounts and can be turned off in Account. Free accounts keep the latest three decks. Paid accounts share a 100 MB allowance with saved notes; when a new deck does not fit, the oldest cloud decks are removed first and notes are never removed by that rule. Turning sync off stops new uploads and leaves existing decks in place until you delete them. Deleting a deck removes its stored archive, and deleting the account removes every stored deck.

Export to Google Slides

If you choose Google Slides as an export, Google asks you to let Video2Any create a file in your Drive. The deck then goes straight from your browser to your Google Drive. It does not pass through our servers, and we do not keep the access Google grants for it.

The browser extension

The Video2Any extension turns the video already playing in a tab into slides. It holds standing access to no website. When you click its toolbar icon, Chrome grants it access to that one tab for that one moment, and it reads still frames from the video element on the page.

No video file is fetched, assembled, or saved by the extension for Mode 1. Frames travel from the tab with the video to the video2any.com tab in the same browser. The extension itself makes no analytics requests. If you then start Mode 2 on that deck, Mode 2 above applies.

While a capture is running, the extension keeps the tab id, the page title, the video's length, and the hostname in session storage (cleared when you close the browser). Nothing is kept about pages you did not click the icon on.

What we collect, and why

Account data. Creating an account stores your email address and a display name. We store no password: signing in sends a six-digit code to your address, and the code is kept only until it is used or expires. Signing in with Google instead stores the name, email address, and account identifier that Google provides. Each signed-in session stores the IP address and browser user agent it started from. To stop abuse of the sign-in form, we count recent code requests per email address and per IP address for a short time. Session cookies keep you signed in; we set no advertising or analytics cookies.

Billing. New checkouts are processed by Paddle.com as our merchant of record when that provider is enabled. Stripe remains available for legacy orders and as a configured fallback. Your card details and billing address go to the payment provider, not to us. We store only what you bought, when, the amount, and the order and customer identifiers needed to match a payment to your account. Their handling of payment data is covered by their own privacy policies.

Contact form. A message through the contact page stores the topic, your email address, the message text, your account ID if you are signed in, and your IP address. The IP address is used to rate-limit spam and for nothing else.

Operational logs. The service runs on Cloudflare. Standard infrastructure logs (IP address, request path, timing) exist for reliability and abuse prevention, on Cloudflare's usual terms. Most records below also store the country your network resolves to, which Cloudflare provides.

Product usage. To see which features are used and where they break, the app reports run metadata (source kind, length, container/codec, picture height, extraction time, page count), export format and edit counts, camera/blank-frame set-aside counts (that check runs on your device), and which buttons you click (label and page path — never what you type). For a pasted link or an extension run, the website's hostname only.

Which pages were opened. Each page render adds a row with the path (never the query string), country, a short device label, and, for crawlers, a short bot name.

Where you arrived from. On your first visit, your browser saves the site that linked you here, any campaign tags, and the landing path, and sends them once. Coming back later does not add a row.

Error reports. When something breaks in your browser, we receive the error message, where in our code it happened, the page path, and your browser's full user agent string. At most ten per visit.

The random id. The records above carry a random id (v2a-anon) created on first visit. Clearing site data gives you a new one. If you are signed in, the records also carry your account ID.

What you tell us on purpose. After an export we may ask whether the slides matched the video; other short questions ask why you are not paying, what you will do with a file, or what stopped a payment. You can also write one sentence from Feedback. Answers store run length and slide counts when we have them — not the file itself.

Internal notifications

Some events send us an internal notification through a messaging service that only the people who run Video2Any read: new accounts (email, name, sign-in method, arrival source, prior conversion count), contact messages, feedback (email shortened), purchases/refunds/chargebacks, and a daily usage summary.

What we do not do

No third-party analytics scripts, no advertising networks, no tracking pixels, no fingerprinting, and no selling or renting of any data. The usage records above are collected by our own code and stored on our own database.

Google

Google's script. If you are not signed in, Google's sign-in script is loaded from accounts.google.com when the sign-in form opens, and in the editor at moments where an account is offered. Loading it sends Google your IP address and the usual request details even if you do not sign in. It is not loaded when you merely visit the site or open the editor.

Google user data. If you sign in with Google, we receive your name, email address, and Google account identifier, and we use them only to create and authenticate your Video2Any account. If you export to Google Slides, we ask for access only to files Video2Any creates in your Drive (drive.file). That access stays in your browser. We do not share, transfer, or sell Google user data. You can revoke access at myaccount.google.com/permissions.

Stored on your device

Your browser's storage holds preferences, current decks and videos, and a few markers so the app does not ask the same thing twice. Two items are sent to us: the random id and where you first arrived from. Nothing else is sent, apart from what you send when you start Mode 2 or when library sync is on. Clearing browser data removes all of it.

Retention and deletion

Account data is kept while your account exists. Detailed usage records (counts, clicks, page views and error reports) are deleted after about 13 months; totals computed from them may be kept longer without any id attached. Contact messages are kept as long as needed to reply and to keep rate-limiting honest.

You can delete your account yourself from the Account section of the app. Deletion asks you to type the account email, then removes the account, its sessions, its OAuth sign-in rows, its saved notes and transcripts, its working Mode 2 files, its shares, stored library decks and settings rows, its API keys and jobs, and the usage and telemetry records tied to the account id. We keep records of completed purchases — transaction and order identifiers, amounts, currencies and dates — without the account profile, so tax, accounting and payment reconciliation stay correct; these may retain an orphaned internal account id. You can still contact us about a deletion through the contact page.

Mode 2 working files are deleted when the job succeeds, or when a stuck job is swept after about two hours. Saved notes and transcripts, and the slide images kept with them for a handout or its share link, have no automatic time expiry; delete them in the app, or delete the account to remove them all.

Who else sees data

  • Cloudflare: hosting, database, file storage, sign-in emails and the sign-in captcha.
  • Third-party AI service providers: audio chunks of a Mode 2 job, uploaded from our servers, for transcription; and, for notes jobs with pages, the transcript text and page images, to write the notes.
  • Paddle and, for legacy or fallback checkouts, Stripe: payments.
  • A team messaging service: our internal notifications.
  • Google: its sign-in script at the moments listed above, and Google sign-in if you choose it.

Each receives the minimum described here. Beyond that, we disclose data only if the law requires it.

Children

Video2Any is not directed at children under 13, and we do not knowingly collect their data. If you believe a child has created an account, contact us and we will delete it.

Changes

When this policy changes, the date at the top moves, and material changes are announced in the product before they take effect.

Contact

Questions about this policy go through the contact page or by email at support@video2any.com; a human reads every message.