Privacy Policy
Updated November 18, 2026
Cliperok turns long videos into short vertical clips. For that the service needs your email address and the videos themselves. This page says what we keep, where it lives, who receives it and how to delete it. The Russian version of this policy is the main one; this is its translation.
Summary
- We keep your email, a password hash, your videos and the clips made from them. Payment details are entered at Paddle and never reach us.
- When you sign in with Google, we take only your email address.
- Videos and clips are stored on equipment owned by the service owner. The website runs on a server rented from OVH.
- Transcript text and video frames go to Anthropic's models, which pick the moments and plan the edit.
- Google Analytics counts visits without cookies and sets cookies only with your consent. There are no ads on the site, and we do not sell data.
- How long each video is kept is shown in the app. You can delete a video earlier with a button in the project.
1. Who is responsible for the data
The data controller is the owner of Cliperok. For any question about your data, write to official@kordar.ai.
2. What we keep
Account
The email address you signed up with, the sign-up date and the invitation code the account was created with. Passwords are not stored in plain text: the server keeps only a scrypt hash.
Google account data
If you sign in with Google, Google gives the app a sign-in token. From it we take only your email address and the flag that the address is verified. We do not read or store your name, profile picture or Google account ID. The app does not ask for access to your mail, Drive or any other Google service.
We use the email address to find your account or create a new one. We keep the address and the time of your first Google sign-in, and we do not store the token. An account created through Google has no password. We do not share Google account data with anyone.
Videos and clips
The files you uploaded or that the server downloaded from your link, and their names. Everything the service derives from them is kept with them: the transcript with the timing of every word, the scan report with the moments found, edit options, finished clips, covers and clip notes.
The scan marks where faces are in the frame and who is speaking, so that a clip keeps the speaker in view. It does not identify people. The labels look like “first speaker” and “largest face”, contain no names, and are not compared across videos.
What you write yourself
Your request to the AI edit, your own hook text, clip ratings and comments, the link to a publication, its date and view count, and your brand style: colours, font and logo.
Technical data
The IP address a request came from. The server uses it to count sign-in and sign-up attempts and stop password guessing. Together with the address, the server log records the request path, the browser name and the page you came from. Logs contain no email addresses or passwords.
The sign-in token. It is issued for 30 days and kept in your browser; the server stores only its hash.
Payments
You enter your payment method, country and postcode in the Paddle checkout, and Paddle keeps them. We send Paddle the plan, our internal account and payment numbers, and your account email so it is filled in at checkout. In return we receive and keep the transaction and subscription numbers, the amount, the currency, the payment and subscription status (whether it renews and when it ends) and your Paddle customer number. The app uses them to grant the plan, show your payment history and open the subscription management page. We never see your card number, address or the email you enter at checkout.
3. What is stored in your browser
Cliperok itself sets one cookie for a year: it holds your analytics choice. Google Analytics cookies appear only if you agreed. In browser storage the app keeps the sign-in token, details of an unfinished upload so the file can resume after a dropped connection, the hints you have already seen, your recent clip sets and your chosen currency. Until you close the tab it also keeps your selected edit settings and the address you came from, with utm tags. All of this is erased if you clear the site data in your browser settings.
Analytics
The site and the app use Google Analytics through Google Tag Manager. It counts which pages are opened and which buttons are pressed, where a visitor came from (referring address and utm tags), the device and browser, and the approximate country. Like any server, Google sees your IP address. We do not send it your email, account number, video titles or the texts you write, and we remove video numbers from page addresses. We need this to see where people drop out of uploading or paying and fix that place.
Until you press «Принять» (Accept), Google receives these signals without cookies and does not recognise your browser on the next visit. The legal basis for this is our legitimate interest in counting visits and finding where people get stuck, without knowing who they are. If you object even to these signals, block googletagmanager.com in your browser, for example with a blocker extension. Google Analytics will then not load at all.
If you press «Принять» (Accept), Google sets the _ga cookie and recognises your browser on later visits. The legal basis for this is your consent. You can change your mind at any time under «Настройки cookie» (Cookie settings) in the account menu or at the bottom of the page. After you withdraw consent we delete the Google Analytics cookies, and signals go without them again. Advertising cookies and advertising signals are always off, with or without consent.
The Google sign-in button loads from Google's servers and may set its own cookies. The Paddle checkout also loads from Paddle's servers and may set its own cookies.
4. Why we need it
Email and password let you sign in. Videos are needed to find moments in them and make clips. Your request and hook text go into the edit, and ratings and views come back to you as a summary in the «Итоги» (Insights) section. The IP address protects against password guessing and overload.
Data is not used for advertising.
5. Who receives the data
- Anthropic
- Transcript text, video frames and your request to the AI edit go to Claude models through the API of Anthropic (USA). The models pick moments and plan the edit. Under Anthropic's rules, data sent through the API is not used for model training by default. The rest is in its privacy policy.
- OVH
- The service owner rents the server the website runs on from this company. Your requests and uploaded files pass through this server.
- Cloudflare
- The Cloudflare network connects the server with the equipment where videos are processed. Requests and files travel over this channel.
- Only if you sign in with Google. Google then knows you signed in to Cliperok. What we receive from it is described in section 2.
- Google Analytics
- Visit signals: without cookies if you have not consented, with cookies if you have. The data is received by Google Ireland Limited and may also be processed by Google LLC in the USA. Google keeps it for 2 months and does not use it for advertising. How Google handles such data is described in its privacy policy.
- Openverse and Wikimedia Commons
- The AI edit may search open photo libraries for a picture for a clip. Then one word or a short phrase goes to their search, usually the name of the thing the video talks about. Videos, transcripts and account data are not sent there.
- Paddle
- Sells the subscription and token packs and takes the payments. Receives the plan, your account email and everything you enter in the checkout. The checkout loads from Paddle's servers, and Paddle sees your IP address. How Paddle handles data is described in its privacy policy.
- Fontshare and Google Fonts
- The site fonts load from their servers. Your browser contacts them directly, and they see your IP address.
Data is not shared with anyone else. The one exception is a court or public authority requiring it by law.
6. Where the data lives and how long it is kept
Videos, clips and the account database are stored on equipment owned by the service owner, and processing happens there too. The website runs on a server the owner rents from OVH. Requests and files reach that equipment through it.
How long videos and clips are kept is shown in the app: every project and clip has a “kept until” date. At the moment this is 7 days after upload. The service owner has decided to keep files with no time limit until you delete them yourself; this section will be updated when that is in place.
The account is kept until you ask us to delete it. The sign-in token is valid for 30 days. Payment records remain after the account is deleted because they are needed for accounting. An unfinished upload is erased one day after the last part was received.
7. How the data is protected
The connection to the site uses HTTPS. Each account sees only its own videos and clips. Access to the server and to the processing equipment is held by the service owner and by those who maintain the service on the owner's behalf.
A link to a clip or a source file is signed and works for one hour. During that hour it opens without signing in, so do not forward it to anyone you do not want to see the file.
8. How to delete data
You delete videos yourself: open the project and press «Удалить видео» (Delete video). The scan report, edit options and all clips from that video are deleted with it. We delete the whole account when you ask us by email. The steps are on the refund and data deletion page.
9. Your rights
Write to official@kordar.ai from your account's email address. We will tell you what data we keep about you and delete the account if you ask.
If you think we handle your data wrongly, write to us first. You also have the right to complain to the data protection authority of your country.
10. Changes
When this policy changes, the new version appears on this page and the date at the top is updated.