Skip to content

Privacy Policy

Last updated

In short. Cevrao finds what is worth filming in your niche, writes a script for it and renders the video. We sign you in with Google and keep what you make: your niches, drafts, uploads and videos. Text goes to OpenAI so a script can be written and read aloud; your uploaded files never leave our server. We do not sell anything, we do not advertise, we do not train models on your work, and nobody records your screen. Ask at any time for a copy of your data or for all of it to be deleted.

1. Who we are

Cevrao is a sole proprietorship (עוסק פטור) registered in Israel, and is the controller of the data described here. This policy covers cevrao.com and everything behind it, not the sites we link to or the platforms you post your videos on. Reach us at contact@cevrao.com.

2. What we collect

  • From Google, when you sign in. Your email address, name, profile picture address and a stable identifier for your Google account. That is all Google tells us.
  • Your answers. Where you post, the niches you follow including one you typed yourself, the parts of them you picked, who you make videos for, the tone, how often you post and how long your videos are.
  • What you do here. Which ideas you see, open, generate from or scroll past, which we turn into a numerical representation of your taste that orders your feed and does nothing else. Your drafts, scripts, edits, caption and voice choices, and your renders.
  • Files you upload. Images and audio you choose. See Uploads.
  • Payment. The fact of a subscription, when it started and what it bought. Card numbers go to Stripe and never to us.
  • Messages. Your name, email and what you wrote, when you use the contact form.
  • Server logs. The address a request came from, the time, the page, the browser string and the response.
  • Analytics, if you accepted them. Page views and clicks. See Cookies.

3. What we do not collect

Worth saying plainly, because the assumption runs the other way.

  • No password. Google is the only sign-in, so there is none of yours here to steal.
  • No access to your Google account. We store no Google tokens, and we cannot see your contacts, Drive, Gmail or YouTube channel.
  • No card numbers. We could not show you your own card if we wanted to.
  • We do not train AI on your work, and neither does the provider we use.
  • Nobody records your screen. Session replay is switched off.
  • No microphone, camera or location. Location embedded in a photo you upload is stripped out.
  • No advertising trackers. There is no ad pixel on this site.
  • We do not read your files. Nobody here opens an upload to see what is in it.

4. Why we use it

  • To run the product - your account, your feed, your drafts, your renders, and the limits your plan actually has. This is performance of our contract with you.
  • To take payment, and to keep the records tax law requires. Contract, and legal obligation.
  • To keep the service up and safe - server logs, abuse guards, diagnosing failures. Our legitimate interest in running a small product securely.
  • To answer you when you write in. Legitimate interest.
  • To see which screens get used. Consent, which you give or refuse in the cookie notice.
  • To answer a copyright complaint, which is why we keep a record of where a placed image came from. Legitimate interest, and legal obligation.

We do not sell your data, share it with advertisers or data brokers, use it to train AI models, or show your drafts and videos to other users.

Your feed is ordered automatically: opening an idea nudges it towards that idea, generating nudges harder, scrolling past nudges away, and behaviour never outweighs what you told us you make. It decides the order of a list and nothing more. Changing your answers in Settings rebuilds it.

5. Your data and AI

OpenAI is the only AI provider we use. We send it text so a script can be written and read aloud: the niche you follow, the text of public articles we fetched, short pieces of text turned into numbers so the feed can be ranked, and the finished script. It reaches OpenAI in the United States, through its GPT, embedding and text-to-speech models. Your name and email are not part of any of it.

Your uploaded images and audio never leave our server. They are decoded, checked and composited into the video on our own machine, and so are the word timings behind the captions. No AI company receives either.

Nothing you do here trains an AI model. We have no models of our own and no plans for any, and OpenAI does not train its models on data sent through its API. We do not read your drafts to improve the product; what we look at is aggregate, like how long a render took.

Model output can be wrong, including about a real person. If a Cevrao draft says something inaccurate about you, tell us and we will deal with it.

6. Uploads, and the requests your browser makes

Metadata is stripped from every upload. Phone photographs routinely carry the coordinates of where they were taken, and a picture taken at home would otherwise hand us your address and then carry it into a video you publish. Uploads are stored outside anything the web server can reach, under a name we generate, and served back only to you. Files that are documents rather than media are refused.

When a draft offers you the share image from a source article, your browser fetches that thumbnail directly from the publisher, because we do not copy it unless you place it. The same is true of stock thumbnails. So opening a draft tells those sites that somebody looked, the way any website sees a visitor. We send no identifier of yours with it.

For every article image placed in a video we record the article, the publisher, the headline, who placed it and when. That record is kept after the draft is deleted, because a complaint about a video made last month has to be answerable.

7. Cookies

A cookie is a small piece of text your browser keeps and hands back on the next request. We use very few, and the analytics one is only set if you accept it - until you answer, and for good if you decline, the analytics library is never started, so there is nothing to set and nothing to send.

CookieWhat it doesHow long
__Secure-authjs.session-tokenKeeps you signed in. Readable only by the server.30 days
__Host-authjs.csrf-tokenStops another site signing you in or out behind your back.The browser session
__Secure-authjs.callback-urlRemembers where to send you after Google finishes.The browser session
__Secure-authjs.pkce.code_verifier, __Secure-authjs.stateOne-time values proving the sign-in coming back from Google is the one this browser started.15 minutes, during sign-in
cevrao_consentRemembers your answer about analytics, so we ask once.1 year
cevrao_navRemembers whether you folded the side rail down to icons.1 year
ph_<id>_posthogAnalytics only, and only if you accepted. Identifies this browser so two page views count as one visit. A matching entry sits in your browser storage.12 months

Our network provider may also set its own security cookies to tell automated traffic from a person. There is no way to serve the site without those, and they are not used to profile you.

There are no advertising or retargeting cookies, no cross-site tracking, no session replay and no social media embeds. Your browser can block or delete cookies for any site, but blocking ours means you cannot stay signed in.

Changing your answer

You have not answered yet, so nothing is being collected. The card in the corner of the screen is where you answer.

8. Who else sees it

Six companies process data for us, each doing one job, and each with its own policy you can read:

  • Contabo (Germany) hosts the server. It runs the site, the renderer and the database, so it physically holds everything: your account, drafts, uploads and videos. Policy
  • Google (United States) provides Sign in with Google, the only way into the product. Policy
  • OpenAI (United States) provides the models behind the research, the script, the feed ranking and the voice. See Your data and AI. Policy
  • Stripe (United States, Ireland) takes card payments. Your card details go to Stripe and never to us. Policy
  • PostHog (European Union) collects the page views and clicks, if you accepted analytics. Policy
  • Cloudflare handles DNS, certificates and the connection in front of our server, and stores encrypted nightly database backups. Policy

That is the whole list. There is no customer data platform, no CRM, no support desk tool, no error tracking service and no advertising stack. Nothing else has a copy.

Two more see your browser but not your data. Thumbnails from Pexels and Pixabay load from their servers while you choose footage, and an article image candidate loads from the publisher, so those sites see the request the way any website sees a visitor. See Uploads.

Data otherwise leaves only for a binding legal request, to answer a valid rights complaint, or if the business is sold - in which case this policy travels with the data and you will be told.

Where it is

Everything is stored in Germany, inside the European Union. Where personal data reaches a processor in the United States, the transfer relies on the European Commission Standard Contractual Clauses or on that provider participating in the EU-US Data Privacy Framework. Cevrao itself is established in Israel, which the European Commission has recognised as providing an adequate level of protection.

9. How long we keep it

  • Your account: while it is open. Close it yourself in Settings and it is deleted at once, not queued for somebody to get to.
  • Your sign-in session: 30 days, or until you sign out.
  • Drafts and scripts: until you delete them.
  • Rendered videos: the file is removed 30 days after the render, by a maintenance pass that runs every six hours.
  • Uploads: until you delete them, then removed by the next maintenance pass, so within six hours.
  • Image provenance: kept, so an old complaint stays answerable.
  • Payment records: as long as tax law requires, up to 7 years.
  • Server logs: rotated, and no longer than 30 days.
  • Analytics events: no longer than 12 months.
  • Messages you send us: up to 24 months.
  • Database backups: nightly, kept for a short rolling window.

One honest exception, in two places. Deleting a draft or a video removes it from your account and removes the file, but leaves a row saying that something existed and has gone. The free plan counts drafts and videos, and without that row somebody could reset a lifetime allowance by deleting and starting again. It holds no script, no video and no content of any kind.

Closing your account takes those rows with it, so the count is kept one more way: we write down how many drafts and video generations the account had used, against a one-way hash of your email address rather than the address itself. We cannot read an address back out of it and we do not keep a list of who has left; what it can do is recognise the same address if it signs up again, so the free allowance is not handed back. It holds two numbers and a date, and nothing else.

Creating a brand-new niche can start paid research. To keep its daily limit from resetting when requests overlap or an account is recreated, we temporarily record the same one-way address hash, an internal niche reference and the time it was created. Each record is deleted after 24 hours.

10. Your rights

Wherever you live you can ask us for a copy of your data, for anything wrong to be corrected, for it to be deleted, for us to stop using it while a question is settled, or to object to a use we base on legitimate interests. You can withdraw consent you gave. Asking never costs you anything and never changes how we treat you.

Email contact@cevrao.com from the address on your account and say what you want. We reply within 30 days and usually much sooner.

On deletion: the account, your answers, your taste vector, your drafts and scripts, and your uploads and rendered files all go. You do not have to ask - Settings has a Delete my account button and it does all of this the moment you confirm. Files are removed by the next maintenance pass. What survives is what the law requires us to keep, such as payment records, the image provenance above, and the hashed record described under How long we keep it. Deleting your account does not delete videos you have already posted elsewhere - only you can remove those.

If you are in the EEA, the UK or Switzerland

The GDPR and UK GDPR apply and the rights above are your Articles 15 to 22 rights. If we have not put something right you can complain to your national data protection authority, or in the UK to the Information Commissioner. We would rather you came to us first, but the right is yours either way.

If you are in California

You have the right to know, delete, correct, and to opt out of the sale or sharing of your personal information. We do not sell your personal information and we do not share it for cross-context behavioural advertising, have not in the last 12 months, and run no advertising, so there is no Do Not Sell link to give you. The categories we collect are identifiers, commercial information, internet activity, audio and visual content you upload, and inferences used to order your feed. We collect no geolocation, no biometrics and no sensitive personal information. Similar laws in other US states are handled the same way.

Elsewhere

Israel, Brazil, Canada, Australia, New Zealand, India, Japan and South Korea all give equivalent rights, and we answer requests under any of them through the same address. In India a person under 18 needs verifiable consent from a parent or guardian.

11. Children and security

Cevrao is not for children under 13 and we do not knowingly collect anything about one. Where the law where you live sets a higher age for consenting on your own, that age applies. If you are a parent and believe a child in your care has an account, write to us and we will close it and delete what is in it.

Everything is served over HTTPS. Every request that touches your work checks who you are on the server and is scoped to you in the database query itself. Uploads are stored privately under names we generate, decoded and checked rather than trusted, and secrets never reach the browser. Anything that costs real money has a limit enforced on the server.

No system is perfectly secure and we will not pretend this one is. If there is a breach that puts you at risk we will tell you and the relevant authority as quickly as we can, and within 72 hours of becoming aware where the GDPR requires it. To report a security problem, email us with Security in the subject line.

12. Changes, and how to reach us

When what we do with data changes, this page changes with it and the date at the top moves. For a change that materially affects you we give notice by email or in the product first. Every privacy question, request or complaint goes to contact@cevrao.com, or the contact form. A person reads it.

Also on this site: Terms of Service and Copyright and Takedown.