PRIVACY NOTICE · UPDATED 2026-11-03

Privacy Policy

We collect only what is needed to provide previews, paid exports, account access, fraud prevention, and support. We do not sell your personal information or use your media to train AI models.

Who operates the service

RemoveBackgroundMusic operates this service and is responsible for the account, purchase, and usage records described in this notice.

Media processing

When you choose an MP4 file, your browser extracts its audio track. The original video remains on your device; only the extracted audio is uploaded for processing through fal.ai or Replicate. Temporary audio storage may also use Cloudflare R2. MP3 and WAV files are uploaded as audio. For an MP4 export, your browser downloads the processed audio and combines it with the original local video.

File retention and deletion

Result access through this service is available for up to 24 hours. For fal.ai processing, we request 24-hour automatic expiration of uploads and generated audio. For singing removal and any enabled Replicate processing, we also retain temporary audio in private Cloudflare R2 for the service recovery window, with an automatic expiration policy. Providers manage physical deletion; this service does not independently confirm its completion. Download results promptly. We do not keep a server copy of your original video.

After your browser combines processed audio with your video, it may retain the completed video in browser storage for up to 24 hours to support a direct re-download on that browser. Browser storage is not a guaranteed backup: clearing site data, closing a private browsing session, browser storage eviction, or changing browser or device can remove it. If the processed audio is still active, you can select the original video again to rebuild the result locally. Download completed videos promptly.

Free previews and abuse prevention

Free previews are limited to one preview per day, covering at most half the file: up to 15 seconds for audio and video, subject to a site-wide daily free-processing budget. To enforce that limit and prevent concurrent abuse, we create one-way identifiers from the request IP address, a random browser identifier, the date, and a server secret. These daily identifiers are automatically removed after no more than three days; the raw browser identifier stays in your browser. Site-wide budget reservations contain random reservation identifiers, the UTC date, a reserved amount and timestamp, not your media or account details. Standard infrastructure logs can include IP address, requested URL, timestamp, and user agent.

New minute packs use their minute balance for previews and full processing. Older purchases retain any included preview counts originally granted. We store the source file fingerprint with the processing record so a successful preview of the same file and selected time range can be reused for up to 24 hours without sending the audio for processing again. The fingerprint is a one-way identifier, not a copy of the media.

Google sign-in

Google sign-in is required to purchase and use paid processing. Google provides a stable account identifier, verified email address, display name, and—when available—profile picture, given and family names, locale, and organization domain. We record login timestamps and counts so your processing history, purchases, and remaining minutes stay associated with the correct account. We never receive your Google password.

Payments

Stripe is our default checkout provider, and Creem may be used as a backup checkout provider. The provider shown at checkout processes the payment and billing details you enter. We store its purchase, transaction and customer identifiers, the selected plan, payment status, entitlement balance, expiry date, and entitlement events. We do not receive or store full card details.

Processing and account records

We retain operational records such as job identifiers, file fingerprints, duration, status, purchase records, entitlement reservations and restorations, and security events. A fingerprint is a one-way technical identifier, not a copy of your media. These records are used for account history, billing reconciliation, fraud prevention, and support.

For Single File, we temporarily receive the complete extracted audio to validate its duration and record its SHA-256 digest for task binding. That validation upload is not retained as a media file. The task identifier, digest and duration remain operational account records; the original video stays on your device.

Analytics and cookie choices

When you accept optional analytics, first-party funnel events use random browser, session and visit identifiers and task and checkout-attempt identifiers. After verified sign-in, the browser identifier may be linked to your account to reconcile the same journey. Client event records are distinct from authoritative server and payment-provider events. Funnel event records are deleted after 90 days and inactive identity links after 180 days by scheduled maintenance. Declining optional analytics prevents these browser identifiers and client funnel events; essential payment and delivery records are still retained.

Usage diagnostics include file format and codec categories, file-size ranges, media and preview durations, preview playback coverage, processing stages, error codes, and checkout return outcomes. A random session identifier links these stages within a browser session and may be attached to checkout metadata for reconciliation. Analytics events do not include your file name, media contents, email address, or payment card details. Playback coverage excludes seeking and does not establish whether a person listened attentively.

We use Plausible to understand aggregate site usage without analytics cookies. We also use Google Analytics with analytics storage denied by default, which permits anonymous measurement until you choose to allow optional analytics cookies. Accepting or declining optional cookies does not affect Plausible or your ability to use the service. Your choice is stored in this browser and can be reset by clearing site data.

Feedback and support

If you submit feedback, we store your chosen topic, message, page path, browser family, timestamp and optional reply email privately for support and product improvement. Do not include sensitive data or original media. We use a daily one-way IP-derived identifier to limit spam; the feedback record does not store your raw IP address. Feedback is not published as a review, and its message and email are not sent to analytics. Contact support to request deletion.

Optional preview-quality feedback records the selected track, processed playback coverage and mute state, issue type and moment, and processing configuration. Missing playback observations remain unknown. Only if you explicitly agree may support listen to your processed preview before its existing expiry. This permission does not extend media retention.

Optional full-result feedback stores your selected save/open status, sound-quality answer and issue category, linked to your signed-in account and paid processing task, with submission/update times and the task’s delivery status. You can change your answer. These answers are internal product feedback, not public reviews, and are not sent to analytics. Submitting an answer does not upload a file, authorize staff to listen to your media, or change your purchase or processing credits. Contact support to request deletion.

Your choices

You may use a preview without creating an account. You can ask for access to or deletion of account information, subject to records we must keep for security, payment, tax, dispute, or legal obligations. Only upload media you are authorized to process.

Contact

For privacy or deletion requests, email support@removebackgroundmusic.com or use the contact page. Do not attach original media to a support message.