Your data,
handled honestly.

• Account email: via Google OAuth through NextAuth.
• Onboarding answers: your role, content and editing preferences, optional public channel profiles, and how you heard about us.
• Uploaded clips and audio you provide to the editor.
• Auto-generated transcripts produced from your clips by our AI pipeline.
• Final exported MP4s rendered through our cloud render pipeline.
• Usage data: analyses run, exports made, credits consumed (used for billing and quota enforcement).
• Error logs via Sentry with PII scrubbed (no email, no IP, no session tokens).
• Person-segmentation mattes: for the "text behind you" effect we run automated person-segmentation (an outline of the person vs the background) on your video. This is not facial recognition — no face template, faceprint, or identity match is ever created, and the matte is deleted with your project.
• Reference reels (Edit like this): If you add a reference reel, we read its editing style once with an automated model (plus one short read of its motion), keep labels and numbers that describe that style, and delete the file. It never appears in your reel. We keep a few still frames from it for about a day while we read its style, and small stills of any graphic in it we couldn't match yet, which we use to improve our template library and delete 30 days after we review them. The file is deleted as soon as it is read; a storage rule removes one that was never read within a day. Your reel gets music from our own library, never the reference's song.
• Connected assistants (Claude, ChatGPT): if you connect one, we keep a connection record (which assistant, when, and what it was allowed to do) and hashed sign-in tokens. The assistant only receives what its tools return: project titles and status, timeline captions and segment times, credit amounts, editor links, and the preview frames you pay for, drawn from your own timeline. Stock-clip search results (public stock descriptions and preview pictures) are tool results too. Where the assistant shows the EditorOP editor inside the chat, your reel's media plays in the editor panel through short-lived private links; those links go to the editor panel, not to the assistant's model. What you type into the assistant stays with that assistant's provider under their terms. When you disconnect it in the assistant's settings, the connection stops working at once. We keep the record of it until you delete your account.

We share only what is required for each service to function. We never sell your data.

• OpenAI / Groq: speech-to-text transcription of your clips' audio. Under their API terms neither uses your content to train models; OpenAI retains transcripts up to 30 days for abuse monitoring, then deletes them.
• AWS: AI models that make editing decisions (the cut, social copy, and asset mapping), running inside AWS. AWS runs these models in accounts the model providers cannot access — per AWS, the model providers never see your prompts or transcripts, and AWS does not use this content to improve its services.
• OpenRouter: routes AI editing calls — cut, hook, graphics, and b-roll placement — only to providers configured not to store or train on your data. The AI director reads your transcript and rendered preview frames of your reel (which can include your face). Those providers are configured with data collection denied: they do not store or train on your prompts. The same rule covers the one-time read of a reference reel: a sheet of still frames from it goes to those providers, and only style labels and numbers come back.
• TypeSafe: short classification questions about your transcript text — labelling each line, spotting a repeated take, judging whether a stock clip fits a line, recognising a named person, brand, or place — and the prompts you type in the Generate panel. It receives text only, never your video or images.
• Cloudflare R2: encrypted storage for your video assets.
• Vercel: frontend hosting.
• Railway: backend hosting.
• Supabase: account database.
• Dodo Payments: payments and tax. Dodo is the seller of record on your invoice and handles VAT/GST/sales tax globally.
• Sentry: error monitoring (PII-scrubbed).
• Resend: email delivery — the receipts, failure notices and, if you opt in, product news. Holds your address and the delivery status of what we sent you.
• Modal: GPU compute for the behind-the-speaker cutout. Receives the seconds of footage being cut out, nothing else, and keeps nothing after the job.
• Unsplash: imagery for concept graphics. Receives a search word derived from your transcript; no other personal data is shared.
• Vercel Analytics & Speed Insights: privacy-friendly, cookieless traffic and performance metrics. Aggregated; no cross-site tracking.
• PostHog: product analytics, so we can see which features work and where people get stuck. For visitors who are not signed in it runs cookielessly: nothing is stored on your device and no profile is built. Once you sign in, your usage is linked to your account, because running the product you paid for is why we measure it. Screen recordings are covered below: always on once you are signed in, and for signed-out visitors either on by default with an opt-out or, in the EU and UK, only if you say yes.
• Google Analytics: traffic measurement: which pages people visit and where they came from. It sets first-party cookies (see Cookies) and follows the same rule as screen recordings: on once you are signed in, and for signed-out visitors either on by default with an opt-out or, in the EU and UK, only if you say yes. Google signals and ad personalisation are turned off, so it builds no advertising profile.
• DataFast: which sites and campaigns bring visitors, and which of them go on to sign up or buy. It sets a first-party cookie (see Cookies) and follows the same rule as Google Analytics. It receives page views and the same product events as PostHog, such as starting an export, but never what you typed or the content of your footage. When you buy, DataFast also receives the payment and subscription details from our payment provider so revenue can be tied to where you came from.
• Pexels, Giphy, Brandfetch, Wikipedia: stock footage, sticker, and logo lookups. Search terms derived from your transcript or typed by you are sent to find matching assets; no other personal data is shared. Assets they return are royalty-free / free to use; see our Terms for how stock assets and your own uploads are licensed.
• What we never do: we never sell your data, and we never use your videos, audio, or transcripts to train AI models — ours or anyone else's. If that ever changed it would be prospective and opt-in only, never applied to content you already uploaded.
• Anthropic / OpenAI: only when you connect Claude or ChatGPT to your account, and only the tool results listed above under connected assistants.

• Clips and projects: kept for 30 days after your last edit on the Creator plan, and 7 days on every other plan. We email you before a project is deleted, and editing it starts the clock again. Deleted sooner if you delete the project or your account.
• Finished videos: downloadable for 7 days after you export. Download a copy to keep it.
• Transcripts at our AI providers: OpenAI retains API inputs for up to 30 days for abuse monitoring, then deletes them. AWS does not retain your prompts or outputs. OpenRouter routes AI editing calls only to providers configured to deny data collection, so they do not store or train on your prompts or frames.
• Account email and billing records: kept while your account exists. Financial records are retained for up to 7 years per applicable tax law, even after account deletion.
• Sentry logs: 90-day rolling retention.
• Backups: deleted data can persist in encrypted backups until they are overwritten on our normal rotation schedule. Backup copies are never restored to re-instate deleted data, and are used for disaster recovery only.

We aim to honor GDPR- and CCPA-style data rights for all users globally, regardless of jurisdiction. We do our best in good faith.

• Legal basis (EEA / UK): we process your data to perform our contract with you (running the Service), on your consent (where we ask for it, e.g. optional analytics), and our legitimate interests (security, abuse prevention, and improving the product), balanced against your rights.
• Access & portability: email privacy@editorop.com from your account email and we'll send a machine-readable JSON export of your account, consents, projects, and library within 30 days. Your finished videos download as MP4s from the editor at any time.
• Deletion: email privacy@editorop.com from your account email and we'll delete your account within 30 days. This removes your videos, exports, projects, and library. Billing records are retained per tax law; deleted data can persist in encrypted backups until rotation (see Retention).
• Correction: to correct your account email, contact privacy@editorop.com.
• Object, restrict & withdraw consent: you can object to or restrict processing, and withdraw any consent you gave, at any time.
• California (CCPA / CPRA / CalOPPA): you have the right to know, delete, and correct your personal information, and to opt out of its "sale" or "sharing." We do not sell or share your personal information, and we will never discriminate against you for exercising any privacy right. Do Not Track: recordings, Google Analytics and DataFast follow the choice described in Cookies; we do not respond to browser DNT signals separately.
• Children: EditorOP is for adults (18+). We do not knowingly collect data from anyone under 18; tell us if you believe a minor signed up and we will delete it.
• Complaints: EEA / UK users may lodge a complaint with their local data protection authority. We'd appreciate the chance to put things right first.

• NextAuth session cookies: required for sign-in. HTTP-only, secure, SameSite.
• Sentry error context: anonymized debugging session ID.
• Measurement (no cookies): PostHog and Vercel Analytics count pages and clicks without storing anything on your device. No cookie, no identifier that survives the tab, no cross-site profile.
• Where you came from: the campaign tags and referring site of your first visit, so we can tell which channels bring people who sign up. It is kept in your browser tab only, and saved in your browser once you sign in or recordings are allowed; choosing Not now or Opt out moves it back to the tab. It is linked to your account when you sign in.
• Google Analytics cookies: first-party cookies (_ga and _ga_<id>, kept up to 2 years) that tell a returning visit from a new one. They follow the same rule as session recordings, below: choosing Not now or Opt out stops Google Analytics and clears them.
• DataFast cookie: a first-party cookie (datafast_visitor_id) that tells a returning visit from a new one. It follows the same rule as Google Analytics: choosing Not now or Opt out stops it.
• Session recordings: a replay of your own screen on EditorOP, which is how we find what breaks. Signed in, sessions are always recorded: you accept this when you sign in, as part of these terms. Signed out, it depends on where you are. In the EU and UK (judged from your device's time zone) nothing is recorded until you choose Allow recordings. Elsewhere recording is on by default and the banner's Opt out turns it off. Either way, choosing Not now or Opt out, then or later, stops it and clears what was stored. We use no advertising or cross-site-tracking cookies, and we never sell your data.
• Essential sign-in and security cookies always remain. You can change your analytics choice at any time via the Manage cookies control — withdrawing is as easy as accepting.

EditorOP is operated from India; for data-protection purposes the data controller is EditorOP (contact privacy@editorop.com). Data is processed in the United States and India (AWS, OpenAI, OpenRouter, TypeSafe, Sentry, Cloudflare, and our hosting). Where we pass personal data to sub-processors outside the EEA / UK, we rely on their Standard Contractual Clauses and, where certified, the EU-US Data Privacy Framework. EEA / UK users can lodge a complaint with their local supervisory authority at any time — we'd appreciate the chance to put things right first at privacy@editorop.com.

We'll notify users by email for any material changes, with at least 30 days notice. Continued use of the Service after the effective date constitutes acceptance.

For privacy questions or data requests, email privacy@editorop.com. For security issues, see our Security page.