Last updated: November 17th, 2026
Privacy Policy
This Privacy Policy explains how ReelBolt ("we", "us", or "our") collects, uses, discloses, and protects information in connection with the ReelBolt public marketing website and the hosted ReelBolt Service. Please read it carefully. If you have questions, contact us using the details in Section 14.
1. Who we are
The ReelBolt website and Service are operated by Tommaso Celano, who is the data controller.
ReelBolt is operated by Tommaso Celano, an individual trader, at Via Fratelli Castaldi 7, 28060 Cureggio (NO), Italy, P.IVA IT02862070030.
Privacy enquiries: [email protected]
We have not appointed a Data Protection Officer: at our scale we are not required to appoint one under Article 37 GDPR, and data-protection questions go to the privacy contact below.
2. Scope
This policy covers two things: (a) the public marketing website you are currently reading — the pages served at the root domain, such as the homepage, features, pricing, about, and contact pages — and (b) the hosted ReelBolt Service and the accounts you use to sign in to it. The website handles contact-form submissions, site analytics and the short links that lead to it. The Service processes account, billing and usage data, and the project files you upload, to do the work you ask of it.
3. What we collect
We collect the following categories of data:
- Contact-form submissions. When you use the contact form, we collect the name, email address, company name (optional), and message you provide, and whether you ticked the optional box to receive our marketing email.
- Marketing preferences. If you ask to receive our marketing email, from the contact form, your account settings, checkout or anywhere else we offer it, we record your email address, where you asked, the exact wording you agreed to and when. If you confirm by clicking the link in our confirmation email, we also record when you clicked and the IP address you clicked from. If you later unsubscribe, we record when.
- Server access logs. Like most websites, our hosting infrastructure automatically records technical information about each request, including your IP address, user-agent string, the requested page address (without anything after a question mark), the page that referred you, the response status and a timestamp.
- Short links. Some of the links we post lead to the website through a short address on our own domain: /l/ followed by the link's name. When such a link is opened, we add one to a counter for that link and that day, and then send the browser on to the page the link points to. The counter is only a number. It holds no IP address, no cookie, no visitor ID and nothing else about the person who opened the link, and nothing is stored on or read from your device. We read the user-agent string of the request and the browser's request-type headers once, to tell a link preview fetched by an app, a crawler, or a page a browser loads ahead of time from a person opening the link. The request-type headers are a few short words in which a browser says whether an address was opened as a page or fetched in the background. The two kinds of request are counted separately, and neither the user-agent string nor those headers are kept with the count. The request reaches our servers through Cloudflare, the provider that serves the website (Section 6). It receives the IP address and the user-agent string of the request, as it does for every page of the website, and limits how many short links one address can open in a short time. The request itself is written to the server access logs described above, like any other request. The page you are sent to has campaign labels in its address, which the entries on visit and sign-up source data below describe. If the click cannot be counted at that moment, you are sent to the home page with the link's name as its one campaign label.
- Analytics data. Our website uses Google Analytics. Before you make a choice in the cookie banner, Google Analytics runs in a cookieless mode. It sets no cookie and stores nothing on your device. It receives a basic page-visit signal: the page address and title, the page that referred you, your browser language, your screen size, your browser and operating system details, and your IP address, from which Google derives an approximate location. The page address includes any campaign labels in the link you followed. The signal carries a random number that is new for every page you open. Besides page views, Google Analytics also records some interactions on the page automatically, in the same cookieless way before you choose: the time you spend on a page, scrolling to the bottom of it, clicks on links that lead off our website, searches made on the site, file downloads, video engagement, and that a form was started or submitted (not what you typed). If you reject cookies, Google Analytics is not loaded and receives nothing more from your visit (the page you were on when you rejected may already have sent its page-visit signal). Only if you affirmatively accept cookies via the cookie banner does Google Analytics set its cookies and collect fuller usage data such as pages viewed, session duration, device/browser type, and approximate location derived from your IP address. We run no advertising, and Google Analytics advertising features stay switched off. The same choice also applies to the ReelBolt app: the app loads Google Analytics only once you accept, with Google signals and ad personalization switched off, and with identifying parts of the page address removed before anything is sent.
- Visit and sign-up source data. The address you arrive on may carry campaign labels. They are short words we put in our own links to say where a link was posted: utm_source, utm_medium, utm_campaign, utm_content and utm_term (on a short link, utm_content is the link's name). They describe the link, not the person who follows it. Only if you accept cookies in the cookie banner, the website gives your browser a random visitor ID (stored in your browser's local storage as rf-visitor-id) and a first-party cookie called rb_ft. The cookie remembers where you came from: the campaign labels in the link you followed, the website that referred you, the first page you landed on, your visitor ID and the time. It lasts 30 days. The app's sign-up page sets the same cookie, for the browser session only, when you accepted cookies, no such cookie exists yet and the address of the sign-up page carries campaign labels. The cookie and the visitor ID belong together: on the website, when the cookie is gone and you visit again, a new cookie is set together with a new random visitor ID, and the same happens when the sign-up page sets the cookie. There is one exception. If you open the app's sign-up page while no cookie exists and its address carries no campaign labels, no new cookie is set, and the two sign-up records described next are sent with the visitor ID already stored in your browser, which may be the one an earlier visit or an earlier sign-up used. When visit recording is switched on, and still only if you accepted, the website and the sign-up page send us short records: that a visit started, which page of the website you opened, which call-to-action button you pressed (such as Start free), and that you opened the sign-up form, sent it, or pressed the Google or GitHub button on it. A page is recorded as one name from a fixed list (for example the home page, the pricing page, or one named blog post), and a button as a fixed name for where it sits (for example the header). Each record holds your visitor ID, the kind of record, the time, and that one name or a source label (the campaign source, else the referring website, else the word "direct"). It never holds anything you typed, and never the full page address. Every record is sent together with the first-touch details the cookie holds (the campaign labels of the link you first arrived on, the website that referred you and the path of the first page you landed on), and the first record we receive stores them once against your visitor ID as your first touch. These records hold no IP address and no user-agent string, although the requests that carry them are also written to the server access logs described above. When you create an account, we store the same first-touch information against it (source label, campaign labels, referring website, first page and visitor ID), so that we can tell which channels bring sign-ups, and the cookie is removed when you sign up. If you then go back on the sign-up page to correct your email address, the cookie is put back until you close the browser, so that the account you really use carries the same first touch. If you never accepted cookies, nothing described in this entry is stored. The next entry describes what can still be saved when you sign up.
- What a sign-up link carries. The sign-up buttons on the website add up to three things to the address of the sign-up page. One is an audience hint, when you came through a page or a link made for one audience (for example for=creator). Another is the campaign labels of the link you arrived on. The third is the landing path: the path of the page of the website your visit began on (for example lp=/pricing). It is added only together with campaign labels, and it is the path alone: nothing that followed a question mark in the address, and never the website that referred you. A sign-up button on the pricing page also names the plan you chose (for example plan=creator), so that the app can take you to that plan once your account exists; the plan is not saved with your account when you sign up (Section 5 describes where the app keeps it in the meantime). Nothing on your device is used for this: the values travel in the link you click and in the sign-up request, whether or not you accepted cookies. The audience hint is saved with your new account as a suggestion, so that the welcome questions described below can start on the matching answer. The campaign labels in the link you followed may be saved with your account when you sign up; they say which of our posts or pages you came from, not who you are. With them we save the time and the landing path the link carried; when the link carried none and the sign-up page was the first page you opened, we save that page's path. If you accepted cookies, the first-touch cookie described above is used for this instead. If you sign up with Google or GitHub, the audience hint and, where we save them, the campaign labels and the landing path pass through that provider's sign-in address as part of a state value, which the provider sends back to us unchanged. Besides a random number, the state value holds only those short words and that path. If a sign-in with Google or GitHub does not complete (for example you cancel it, or no account can be created), you are sent back to the app's sign-in page, and the same values come back in the address of that page next to a short word for what went wrong: the audience hint and, where we save them, the campaign labels and the landing path. This way they are not lost if you then sign up another way. The Sign in link on the website carries the same things as the sign-up buttons, because signing in with Google or GitHub creates an account when you have none. Two kinds of link between pages of the website carry them as well, in the address only: the language links in the footer, and a link from a translated page to a page that exists only in English. Those links open the next page as a new document, and without the values in its address that page would no longer know which link you arrived on or where your visit began. Nothing is stored for this either.
- Account data. When you create an account, we collect the name, email address and password you provide, and the workspaces you create or join.
- Welcome questions. After you sign up we ask two questions: what you are here to make a video for (for example your channel, your business, your lessons, your cause, your software product or your clients) and what you have to start from (clips you already shot, an app or website you built, or just an idea). Each answer is one choice from a fixed list, with nothing to type. Answering is optional, and you can skip the questions. We store your answers, or that you skipped, and when, with your account, next to the audience hint from your sign-up link if there was one. We use them to choose your first recipe and the examples you see, and to count in our own statistics what kinds of people sign up and how far they get. Those statistics are produced with the reporting tool named in Section 6, from a nightly copy that holds your answers next to your account ID. They are deleted with your account.
- Billing data. When you subscribe or buy credits, payment is handled by our payment provider (currently Stripe and Paddle, see Section 6). We receive the subscription or transaction status we need to run your account; we do not receive or store your full payment card details.
- Usage data. We meter how the Service is used — for example which runs you start and the credits they consume — so that we can operate and bill the Service.
- First-step records. We record that certain steps happened on your account, most of them only the first time: for example that you created your first project, uploaded your first file, downloaded your first video, and answered or skipped the welcome questions. Each record holds the name of the step, the time, your account and workspace and, for some steps, one fixed word, such as your answer to the first welcome question or the recipe you picked. It holds no file, file name, text or other content. These records are written by our own servers when the step happens, not sent by your browser, so they do not depend on the choice you make about the funnel analytics described below. They are part of the nightly copy that our reporting tool reads (Section 6).
- Uploaded content. The files and text you upload to the Service, which we process to generate your videos.
- Avatars. If you create an avatar of a person or of an object, we keep the reference pictures, short videos and short speech samples you attach to it, the person's name as you type it, whether you are that person or have their written permission, your confirmation that they are 18 or older, a signed release if you choose to upload one, and the consent record: who confirmed, when, the exact statement shown, its version and its language. For an avatar of an object we keep the pictures and the confirmation that they show no person. If an Owner or Admin saves an avatar to the workspace avatar library, it is visible to and usable by the members of that workspace, and we keep that confirmation too. We also keep a log of when the avatar was created, changed, revoked and deleted. The log holds no pictures, recordings, links or the person's name. The person shown may not be a user of the Service.
- Community recipes. If you publish a recipe, the public gallery shows an author label. It starts as your account display name, and you can change it for each recipe. No other detail about you or your workspace is shown. The gallery also shows how many workspaces copied a recipe and, from 5 ratings, the average of its ratings. We store who rated a recipe and who reported one, and we never show that to the author or the public. We keep the versions of a recipe while it is published, and after removal only as long as we need them. We also keep the reports and a moderation log of approvals, rejections, removals and hides. Routine moderation log entries and handled reports are deleted 12 months after they were written or handled. Log entries, and their related trail, tied to a safety, illegal-content, copyright or rejected-by-moderation finding are kept longer, as needed for safety and legal duties. A dismissed report is cut to a minimal record (the recipe and the account only, no reason or note) after 12 months, until the account is deleted.
- Product telemetry. Traces and metrics about how the Service runs: which operations a request performs, how long they take and whether they succeed. They carry account and workspace identifiers, not your files, prompts or other content.
- Error tracking. When something breaks, a record of the failure, on our servers and in your browser. A browser error report carries the error class name and the page it happened on, with identifying parts of the address removed. It does not include the error message or stack, because they can carry your content.
- Funnel analytics. Four events that mark the first time you sign up, create a project, run a workflow and export. Each event carries its name only and is sent only if you allow it.
- The feedback channel. The bug reports and feedback you choose to send us from inside the product: the message you write and any context you choose to attach.
- Ambient product-health signals. Automatic signals that tell us the alpha is struggling: a failed tool call, a refused request, a failed workflow step, or a chat that ended on an error. They are metadata only, with no message content, and are linked to your account and workspace (and so to your email address), which means they are not anonymous.
First-video abuse check. When you create an account, ipinfo.io receives your sign-up IP address for fraud prevention. We may withhold or withdraw the free first video if we detect abuse or multiple accounts.
4. Why we collect it, and our lawful basis
We only process personal data where we have a valid lawful basis to do so. The list below summarizes the purpose and legal basis for each category described in Section 3.
- Contact-form data is processed to respond to your enquiry and, where applicable, to take steps you request prior to entering into a contract with us. Lawful basis: performance of, or steps prior to, a contract, and/or our legitimate interest in responding to enquiries directed at us.
- Server access logs are processed to operate, secure, and troubleshoot the website (for example, detecting abuse or diagnosing errors). Lawful basis: legitimate interest in keeping the website secure and functioning correctly.
- Analytics data is processed to understand how visitors use the site and to improve it. Lawful basis: your consent, obtained through the cookie banner described in Section 5, for Google Analytics cookies and the fuller measurement they allow; and our legitimate interest in knowing whether the website is reached and works (Article 6(1)(f) GDPR) for the cookieless page-visit signal described in Section 3, which Google Analytics receives only before you choose; rejecting cookies ends it. You may withdraw your consent at any time. For questions about the cookieless signal, contact us using the details in Section 14.
- Short-link counts tell us how often each link we post is opened. Counting stores nothing on your device and keeps nothing that identifies a person, so we do not ask for consent for it. Handling the request itself, like any other request to the website, rests on our legitimate interest in running the website and knowing which of our posts are opened (Article 6(1)(f) GDPR).
- Visit and sign-up source data is processed to measure which links, pages and buttons bring sign-ups, so we can decide where to spend our effort. Lawful basis: your consent, obtained through the same cookie banner.
- The audience hint, the campaign labels and the landing path a sign-up link carries are processed for the same purpose, and the audience hint also to start the welcome questions on a fitting answer. They are not read from or stored on your device, so the cookie banner does not cover them. Lawful basis: our legitimate interest in knowing which of our posts and pages bring sign-ups (Article 6(1)(f) GDPR). You can object at any time, as Section 11 explains.
- Account, billing and usage data are processed to provide the Service you signed up for, to bill you, and to keep the Service secure and functioning. Lawful basis: performance of our contract with you, compliance with our accounting obligations, and our legitimate interest in operating and securing the Service.
- Welcome answers are processed to set up your first project and examples to fit what you told us, and to count in our own statistics what kinds of people sign up and how far they get. Lawful basis: performance of our contract with you, for the set-up (Article 6(1)(b) GDPR), and our legitimate interest in understanding who uses the Service, for the statistics (Article 6(1)(f) GDPR).
- First-step records are processed to see where new accounts get stuck, so that we can fix it. Lawful basis: our legitimate interest in operating and improving the Service (Article 6(1)(f) GDPR).
- Uploaded content is processed to carry out the work you ask the Service to do — for example, to generate a video from your files. Lawful basis: performance of our contract with you.
- Avatars are processed only to show the person in AI-generated videos and pictures in the project where the avatar was made, or, if it is saved to the workspace avatar library, in the projects of that workspace, and to keep the proof that they agreed. Pictures of an object are sent to the vision model provider of the workspace, which checks that no person is in them; the avatar is refused if one is found. We treat a person's face, body and voice in an avatar with the care the law requires for sensitive data. Lawful basis: for the person shown, their explicit consent (Article 6(1)(a) and, where Article 9 GDPR applies, Article 9(2)(a) GDPR), which the account holder collects and confirms to us before the avatar can be created; for you as the account holder, performance of our contract with you. The consent record and the log are kept so that we can show when and how consent was confirmed and that the avatar was deleted. Lawful basis: our legitimate interest in proving consent and in defending and responding to complaints (Article 6(1)(f) GDPR) and, where it applies, our legal obligation. A signed release you upload is kept only as that proof and is never sent to any AI provider.
- Marking AI-generated video. Video made with an avatar carries a machine-readable mark that it was generated with AI and, unless the account holder turns it off and discloses it another way, a visible label. We record in the video's details that it was AI-generated and which avatar was used.
- Marketing email (product news, tips and offers) is sent only if you have opted in and confirmed. Lawful basis: your consent (Article 6(1)(a) GDPR). You can withdraw it at any time, as Section 11 explains. Withdrawing does not affect email we sent before you did.
- Records of your marketing choices are kept so we can show when and how you agreed, and so we never email you again after you unsubscribe. Lawful basis: our legitimate interest in proving consent and honoring your opt-out (Article 6(1)(f) GDPR).
- Records of your telemetry choice are kept so we can show whether you granted or declined telemetry. Lawful basis: our legitimate interest in evidencing and honoring your choice, not the consent that governs the telemetry itself (Article 6(1)(f) GDPR). The record is tied to your account and can be linked to your email address, so it is not anonymous.
- Product telemetry, error tracking and product-health signals are processed to run the Service, to find and fix failures, and to see when the alpha is struggling. Lawful basis: our legitimate interest in operating, securing and improving the Service (Article 6(1)(f) GDPR).
- Funnel analytics and browser-side error reports are collected only with your consent (Article 6(1)(a) GDPR), asked for in the product and off until you allow it. You may withdraw at any time, and declining changes nothing about what you can do.
- Feedback you send us is processed to read and act on your report. Lawful basis: your consent when you choose to attach context, and our legitimate interest in responding to what you tell us (Article 6(1)(a) and 6(1)(f) GDPR).
Safety screening. We automatically screen files you upload and images, audio and videos you create against our content rules, to prevent illegal and harmful content. This also covers videos you never open or share. For this we send up to 20 downscaled frames and the text of a file or a transcript to OpenAI or Azure OpenAI. To make the transcript, we send up to 10 minutes of the audio to a transcription provider. The original video file is not sent. A result is stored with the file. Lawful basis: our legitimate interest in keeping the Service safe and lawful (Article 6(1)(f) GDPR). Pictures, videos and audio attached to an avatar are screened before the avatar can be used, and an avatar cannot be created from a file that has not been screened yet.
Community recipes (Section 3). The text of a recipe you publish is checked by OpenAI's moderation service, one of the providers named for safety screening above, and the preview video is screened and scanned for illegal material like other videos. Lawful basis: our legitimate interest in running and keeping safe a public gallery (Article 6(1)(f) GDPR), and, for publishing the recipe itself, performance of our contract with you (Article 6(1)(b) GDPR).
Abuse prevention for the free first video (Section 3). Lawful basis: our legitimate interest in preventing abuse of the free offer (Article 6(1)(f) GDPR).
The payment provider's checkout may offer its own box for receiving the provider's marketing email. That choice belongs to the provider. We do not receive it as permission to send you ReelBolt marketing email, and we never use it for that.
5. Cookies and similar technologies
We use a small number of storage mechanisms, summarized below. We set no analytics or measurement cookie before you make a choice in the cookie banner. Your initial choice itself is remembered in your browser's local storage, not a cookie. A few entries do not depend on that choice, because they only remember something you did yourself: the language you picked, a notice or offer bar you closed, and where you were going when you started to sign up. The table says for each one when it is written. One thing does run before you choose and is not a cookie: the Google Analytics tag loads on every page of the website and sends the cookieless page-visit signal described in Section 3. It stops once you reject. Opening one of our short links stores nothing on your device. Following a sign-up link that carries an audience hint or campaign labels stores nothing on your device either, unless you accepted cookies: then the app's sign-up page may set the rb_ft cookie and the visitor ID from the campaign labels in its address, as the table below and Section 3 describe.
| Name | Type | Purpose | Duration |
|---|---|---|---|
| rf-cookie-consent | Local storage (not a cookie) | Strictly necessary: remembers whether you accepted or rejected analytics cookies, so we don't ask again on every visit. | Persistent, until you clear it or change your choice |
| rb_locale | First-party cookie | Functional: remembers the language you picked, so that the website and the app open in it, and so that emails we send you while no language is saved in your account are written in it. Set only when you pick a language yourself: a language link in the website's footer, or the language menu in the app. It holds the language code and nothing else. | One year |
| reelbolt.alpha-banner.dismissed:<version> | Local storage (not a cookie) | Functional: remembers that you closed the alpha notice at the top of the page, so it stays closed. Written only when you close the notice. It holds a single mark and nothing about you. | Persistent, until you clear it. A new version of the notice is shown once more. |
| reelbolt:promo-banner-dismissed | Session storage (not a cookie) | Functional: remembers which offer bar you closed, so it stays closed in this browser tab. Written only when you close the bar. It holds a short description of that offer (its kind, end date, headline and discount) and nothing about you. | Until you close the tab |
| rb-signup-intent:v1 | Local storage (not a cookie), written by the app's sign-up and sign-in pages | Functional: when the link that brought you to sign up named a plan or a page to return to, remembers those two values so that you arrive there once your account exists. Written only when you send the sign-up form or press the Google or GitHub button. Besides the two values, the entry holds the time it was written and a short code computed in your browser from your email address (not the address itself), so that only that account picks the entry up. For a sign-up by email the code is added when you send the form; for a sign-up with Google or GitHub it is added once you are signed in. The entry is read only by the app in your browser and is not used for measurement. | Removed when used. An entry that is not used lapses after three days at most (after 30 minutes if a Google or GitHub sign-in did not come back in that time) and is deleted the next time you open the app signed in on that browser. |
| rb_ft | First-party cookie | Acquisition measurement: remembers the campaign labels, referring website and first page of your first visit, with your visitor ID and the time. Set only after you accept; removed when you sign up, or when you reject or withdraw. If you go back on the sign-up page to correct your email address, it is put back until you close the browser. | 30 days (a session cookie when the app's sign-up page sets it) |
| rf-visitor-id | Local storage (not a cookie) | Acquisition measurement: a random ID that ties your visits together without naming you. Created only after you accept; removed when you reject or withdraw. | Persistent, until you clear it or withdraw consent. Replaced by a new random ID whenever a new rb_ft cookie is set. |
| rf-visit-sent | Session storage (not a cookie) | Acquisition measurement: marks that this browser tab has already sent its visit record, so that one visit is counted once. Set only after you accept; removed when you reject or withdraw. | Until you close the tab |
| Google Analytics tag | Script (no cookie, nothing stored on your device) | Analytics: sends the cookieless page-visit signal and automatic interaction measurements described in Section 3 while you have not chosen. Not loaded once you reject. | Signals sent while you are on a page |
| _ga | Cookie (Google Analytics) | Analytics: distinguishes unique visitors. Set only after you accept. | Typically 2 years |
| _ga_<container-id> | Cookie (Google Analytics) | Analytics: persists session state for a specific Google Analytics property. Set only after you accept. | Typically 2 years (session cookie variants: 24 hours) |
You can change your mind at any time: reopens the consent banner so you can accept or reject analytics cookies again. If you reject or withdraw, we remove rb_ft, rf-visitor-id and rf-visit-sent and stop Google Analytics from collecting anything further, on later visits too for as long as your choice is stored. Google Analytics cookies that are already in your browser are not deleted by this: they expire on their own schedule, or you can delete them in your browser settings.
6. Sharing and processors
We do not sell your personal data. We share limited data with the service providers below. Each acts on our behalf under appropriate contractual terms, except Stripe and Paddle, as explained after the table.
| Service | Provider | Purpose |
|---|---|---|
| Payment | Stripe and Paddle | Sells subscriptions and credits to you as the merchant of record, as an independent controller |
| Cloud hosting | Hetzner (Falkenstein, Germany) | Hosts the Service and processes its server access logs |
| Website delivery | Cloudflare | Serves the marketing website and its short links, and passes requests for the Service on to our servers. It receives the IP address and the user-agent string of each request and limits how many requests one address can make in a short time |
| Storage | Cloudflare R2 | Stores uploaded content and generated files, and the nightly reporting files described below the table |
| Cloudflare | Delivers contact-form submissions and transactional email | |
| Marketing email | Brevo | Delivers marketing email, only from our own list of people who opted in |
| AI model providers | Black Forest Labs (FLUX), fal.ai (subject detection, depth, image generation and editing, reference-to-video, talking-avatar video and lip-sync), MiniMax, Higgsfield, OpenAI and Azure OpenAI, Anthropic, Google (Gemini), DeepSeek, Fish Audio | Process data sent for ReelBolt-supplied inference when a workflow runs, and for automatic safety screening (OpenAI or Azure OpenAI, and the transcription provider) and the check of published recipe text (OpenAI's moderation service) |
| Telemetry and error monitoring | Grafana Cloud (Grafana Labs) | Stores and displays product telemetry, error reports and product-health signals |
| Internal reporting | Microsoft Power BI | Produces our own statistics from the nightly reporting files described below the table |
| IP reputation | ipinfo.io | Receives your sign-up IP address for fraud prevention on the first-video offer |
| Analytics | Google Analytics (Google Ireland Ltd) | Processes the analytics data described in Section 3. On the website it receives a cookieless page-visit signal before you choose, nothing after you reject, and sets its cookies only once you have given consent; in the app it is loaded only once you have given consent |
Payments are taken by Stripe (Link), which sells to you as the seller of record and merchant of record under Stripe Managed Payments. Subscriptions that began on Paddle are still handled by Paddle, which remains the merchant of record for them. The merchant of record takes the payment, charges any tax and issues the invoice, and decides how it uses the payment and billing details it collects from you, so it is an independent controller of that data under its own privacy notice, not a processor acting for us. We receive only what we need to run your account, such as your subscription status, the transactions on it and any discount applied.
The AI model providers listed above are the model providers you may be routed through when ReelBolt supplies the AI for a run. If you connect your own AI provider account ("bring your own provider"), that traffic goes to your own provider under your own agreement with it — not to the providers listed above. Safety screening is the exception: it always uses the providers we choose.
Product telemetry, error reports and product-health signals (Section 3) are exported to Grafana Cloud (Grafana Labs) in the EU region, which stores and displays them on our behalf under Grafana's data processing agreement.
Internal reporting. Our own statistics, including the sign-up and first-step figures described in Section 3, are produced with Microsoft Power BI. Each night a copy of the records they are built from is written to files held in Cloudflare R2, and the reporting tool reads those files. Each night's files replace the ones from the night before. The copy includes account records (with the name and email address of the account), workspace, billing and usage records, first-step records, welcome answers, the sign-up source stored with an account, and the visit records and first touch of visitors who accepted cookies, with their visitor IDs. It holds no passwords and none of the files you upload or the videos made from them.
6.1 What each AI provider does with your content
Each provider handles the content we send it under its own terms, linked below. These are the same providers the Terms refer to.
| Provider | What is sent | Its terms |
|---|---|---|
| Black Forest Labs (FLUX) | A frame of your footage, or your prompt, when a workflow uses a FLUX model. That includes an image generated or edited on a FLUX endpoint (a thumbnail cover, or a subject moved in the image editor); in that case we send the source image you chose plus the instruction. | May train on your content. Their API terms state that the developer grants a fully paid, royalty-free, perpetual, irrevocable, worldwide, sublicensable licence over inputs and outputs, and that they may use inputs and outputs to train and improve their models. |
| fal.ai (subject detection, depth, image generation and editing, reference-to-video, talking-avatar video and lip-sync) | The image you ask to cut a subject out of, measure depth on, edit or extend, together with what you typed (a subject name such as "the person", an instruction, or a description of what to generate) and, for an erase or a fill, the mask you painted. The image and the mask are fetched by fal itself from short-lived links we create for that one request, never from public addresses. For a talking-avatar video: up to three of the avatar's reference pictures (to make a still of the person) and then that still with the narration audio (to animate it), sent with each request, with a request that fal does not keep them. Avatar and object reference pictures also go to fal for reference-to-video (a clip that shows avatars), for composing a scene picture, for pictures with avatars in the picture editor, for making a variant of an object, and for talking-avatar video and lip-sync; for cutting an object out of a video, stills of your video go to fal, and only stills that show no person. The models fal runs for these are made by other companies, and your content reaches them through fal.ai: Alibaba (Wan) and Google (Veo and the picture-edit model) for reference-to-video, scene pictures and pictures with avatars; MiniMax for lip-sync; Kuaishou (Kling) for a talking-avatar video, when the workspace administrator selects that model; Meta (Segment Anything) for cutting an object out of a video. | |
| MiniMax | A frame of your footage, a reference image, or your prompt, when a workflow uses a MiniMax model. That includes image generation and image-to-image editing through their image endpoint, where the reference image is fetched by the vendor from a link we give it. For a video that uses an avatar: the avatar's reference pictures, short reference videos and short speech samples, sent with each request. | |
| Higgsfield | Only when the workspace administrator selects a Higgsfield video model. Your prompt and, for a clip made from a first frame, that starting picture, which can be a scene picture composed from an avatar or an object, or a picture you chose. We upload the picture to Higgsfield's file storage and send the link. An object variant made from a video also uses the workspace's video model, so it goes the same way when that model is Higgsfield. A Higgsfield model is not used for a clip that shows avatars in reference mode. | |
| OpenAI and Azure OpenAI | A prompt, a transcript, a frame, or a voice recording, depending on which capability a step uses. Also up to 20 downscaled frames and the text or transcript of your uploads and created files, for automatic safety screening, whichever capabilities a step uses. To make a transcript, up to 10 minutes of the audio goes to a transcription provider. Also, to check for a person, each picture of an object and one frame of each video of it, and, when you cut an object out of a video, up to six stills of that video and every frame that was cut out; those stills may show people, which is what the check is for. This applies when the workspace's vision model provider is this one. | |
| Anthropic, Google (Gemini), DeepSeek | A prompt, a transcript, or a frame, depending on which capability a step uses. Also each picture of an object and one frame of each video of it, and, when you cut an object out of a video, up to six stills of that video and every frame that was cut out, to check for a person; those stills may show people, which is what the check is for. This applies when the workspace's vision model provider is one of these. | |
| Fish Audio | The narration text, and for a cloned voice the reference recording you supply. |
7. Google user data
You can sign in to ReelBolt with Google, and you can link a Google Drive folder to a project. This section explains the Google user data involved in both, and what we do with it.
7.1 Sign in with Google
If you sign in with Google we receive your email address, your name and your profile picture, and we use them only to create your account and authenticate you. We store your email address and name with your account; we do not receive your Google password.
7.2 Google Drive sync
When you link a Google Drive folder to a project, we ask Google for read-only access to your Drive (https://www.googleapis.com/auth/drive.readonly), and only to the files and folders you choose to link. We read those files so we can copy them into that ReelBolt project: the sync is a one-way copy, and the files in your Drive are never modified or deleted. Your Drive content is stored with the rest of your project content (Section 6).
7.3 Storage, sharing and retention
The access token we use for Drive is encrypted at rest. We do not sell your Google user data and we do not share it beyond the sub-processors named in Section 6, which process it only to provide the Service to you. A linked folder's content stays in your project until you disconnect the link or delete your account; deleting your account deletes it with the rest of your content (Section 10).
7.4 How to revoke access
You can disconnect a linked Google Drive folder at any time from the project's settings, which stops the sync and removes the stored token. You can also revoke ReelBolt's access to your Google account at any time from your Google account permissions page.
ReelBolt's use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements. Read the Google API Services User Data Policy.
We do not use Google Workspace APIs, or any Google user data received through them, to develop, improve or train non-personalized (generalized) AI or ML models.
8. The desktop runner
If you install and pair the ReelBolt desktop runner, we collect a device identifier, the runner's operating system, and its hardware and software capabilities, so that the Service can send it work it can perform. The runner does not access files on your computer other than the files it is sent for a job. You can unpair a device at any time.
9. International transfers
Some of the processors listed in Section 6 may process data outside of your own country or region. Where that happens, we rely on the following transfer mechanism to ensure your data continues to receive an adequate level of protection: the European Commission Standard Contractual Clauses. This includes ipinfo.io, which receives your sign-up IP address for the first-video abuse check described in Section 3, and the data Google Analytics receives from the website and from the app.
10. Retention
We keep contact-form submissions for 3 years, after which they are deleted or anonymized unless a longer period is required to resolve an ongoing enquiry or comply with a legal obligation. Server access logs are retained for 30 daysfor security and troubleshooting purposes, after which they are automatically purged. Analytics data is retained according to Google Analytics' own configured retention settings for our property. Visit and sign-up records, including the page and button records and the first touch stored against a visitor ID, are deleted after 13 months. The first-touch source we store against your account, whether it came from the cookie or from the campaign labels and landing path in your sign-up link, is kept for as long as the account exists and is deleted with it; after that, any visit records that carried your visitor ID can no longer be linked to you and are deleted within 13 months of the visit. Short-link click counts hold no personal data and are kept as statistics, with no fixed end date.
Your welcome answers, and the audience hint from your sign-up link, are kept for as long as your account exists and are deleted with it. First-step records are kept for 12 months and then deleted.
Account content — your uploaded files and the videos the Service generates from them — is retained for 3 years, after which the account and its content are deleted. You may ask us to remove your content sooner under Section 11.
For marketing email, a sign-up you never confirm is deleted after 30 days. While you are opted in, we keep the record of your consent, and for 3 years after you withdraw it, so we can show what you agreed to and when. After that, we keep only a suppression entry: a keyed hash of your email address and the date you unsubscribed, which lets us make sure we never email you again without storing the address itself. If you delete your account, we delete your marketing records and keep only that suppression entry; if you never opted in, there is nothing to keep. If you later opt in again and confirm, the suppression entry is removed.
An avatar and its files are kept until the date you choose when you create it (between 1 day and 2 years; one year unless you change it), and are then deleted automatically. If you revoke the consent, the avatar can no longer be used from that moment and is deleted 30 days later; you can delete it sooner yourself. Deleting an avatar deletes its reference files and any release you uploaded. The log of when the avatar was created, revoked and deleted, which holds no pictures, recordings or name, is kept for 3 years and then deleted. Videos you made with the avatar are your files and follow the retention of your other content.
Routine moderation log entries and handled reports for community recipes are deleted 12 months after they were written or handled. Open reports are kept until they are handled. Log entries, and their related trail, tied to a safety, illegal-content, copyright or rejected-by-moderation finding are kept longer, as needed for safety and legal duties. A dismissed report is cut to a minimal record (the recipe and the account only, no reason or note) after 12 months, until the account is deleted, so repeated dismissed reports do not count toward hiding a recipe. If you delete your account, we delete your recipes, ratings and reports. The moderation log stays, without your identity. Copies other people made of your recipe stay with them.
We keep product telemetry, error tracking, funnel analytics, the feedback channel and the product-health signals for 12 months, then delete or aggregate them. The record of your telemetry choice is kept while your account exists, and deleted within 12 months after it is closed.
11. Your rights
Depending on where you live, you may have some or all of the following rights over your personal data:
- The right to access the personal data we hold about you
- The right to have inaccurate personal data corrected (rectification)
- The right to have your personal data deleted in certain circumstances (erasure)
- The right to restrict how we process your personal data in certain circumstances
- The right to receive your data in a portable format (data portability)
- The right to object to processing based on legitimate interest
- The right to withdraw consent at any time, for processing based on consent (such as analytics and marketing email)
- The right to lodge a complaint with a supervisory authority, in particular the Garante per la Protezione dei Dati Personali (Italy)
To exercise any of these rights, contact us using the details in Section 14. We will respond in accordance with applicable law.
If you are shown in an avatar. You may not have an account with us. You can withdraw your consent through the account holder who made the avatar, or by writing to [email protected]. When you write to us, we make the avatar unusable while we look into it, tell the account holder, and delete it if consent cannot be confirmed. Copies a provider already received are covered by that provider's own terms (see Section 6.1).
Your right to object to marketing. You can object at any time, free of charge and without giving a reason, to our use of your personal data for direct marketing. Once you do, we stop using it for that purpose. To unsubscribe, click the unsubscribe link in any marketing email (it works in one click, with no need to sign in), turn marketing email off in your account Settings, or write to [email protected]. Unsubscribing takes effect immediately and does not stop the emails you need to use the Service, such as receipts and security notices.
12. Children
The website and the Service are directed at businesses and professionals evaluating or using ReelBolt. They are not directed at, and we do not knowingly collect personal data from, children. If you believe a child has provided us with personal data, please contact us so we can delete it. Avatars and cloned voices of anyone under 18 are not allowed, and the Service asks you to confirm the person's age before an avatar is created. We do not check age ourselves.
13. Changes to this policy
We may update this policy from time to time to reflect changes in our practices or for legal, operational, or regulatory reasons. We will update the "Last updated" date at the top of this page when we do. When a change is material, we will also tell you about it in the Service or by email.
14. How to contact us
If you have questions about this policy or how we handle your data, contact us at:
ReelBolt is operated by Tommaso Celano, an individual trader, at Via Fratelli Castaldi 7, 28060 Cureggio (NO), Italy, P.IVA IT02862070030.
Privacy enquiries: [email protected]
We have not appointed a Data Protection Officer: at our scale we are not required to appoint one under Article 37 GDPR, and data-protection questions go to the privacy contact below.