Privacy & Cookies Policy

This policy explains how ClipShrink handles information when you use clipshrink.app, its account features, and its media, transcription, and translation tools. ClipShrink is a browser-based service; it does not currently provide a downloadable desktop application.

For privacy questions, access or deletion requests, or concerns about a file, email jinjing3333@gmail.com. Include the account email or task identifier needed to locate your request, but do not send passwords, payment-card details, or sensitive media unless support specifically asks for a necessary sample.

Video compression, video trimming and cropping, audio trimming, video-to-audio conversion, and image and PDF compression run in your browser. For these local operations, a file chosen from your device is not uploaded for encoding. Account authentication and credit checks can still send account and job information to our servers.

A public-link import is different: the signed-in import service retrieves the supported public Google Drive, Dropbox, or direct HTTPS source and transfers it to your browser. The link and transfer metadata therefore pass through the service even when the subsequent encoding is local.

Transcription, video translation, and document and image translation use cloud processing. Files or extracted audio are uploaded to private Cloudflare R2 storage and processed by our workers. Extracted text is sent to Groq for translation; transcription audio is sent to Groq or the configured xAI provider. Do not upload material you are not permitted to share with these services.

Account information includes your email, authentication identifiers, and any display name supplied through registration or Google sign-in. Supabase manages authentication and stores account balances, credit movements, and processing history.

Processing information can include file names, sizes, source links, languages, task identifiers, status, errors, timestamps, extracted text, transcripts, subtitle cues, translations, and generated download files. The information handled depends on the tool you choose. We use it to provide the requested result, show your history, calculate usage, restore failed-task credits, and troubleshoot processing.

For document and image translation, a signed browser-session identifier and hashed network identifier enforce the free allowance. Cloudflare Turnstile receives verification information, including the IP address used for the challenge, to help prevent automated abuse. Hosting and worker services can also process technical request and error logs for operation and security.

If you contact support, we use the message, account or task details, and any attachments you choose to provide to respond and investigate the issue. Only share information that is necessary for that request.

When paid checkout is available, payments are handled through Creem as merchant of record. Creem receives checkout, customer, billing, and payment information and applies its own buyer terms and privacy notice. Full payment-card details are entered in Creem’s hosted checkout, not collected by ClipShrink.

ClipShrink stores the account association, purchased pack, price and currency, payment and order identifiers, status, credited amount, and relevant refund or dispute events. We use these records to deliver credits, reconcile payments, resolve support requests, prevent duplicate credit grants, and reverse refunded or disputed purchases. Credits are not delivered solely because the browser returns from checkout.

Purchases are unavailable while payment setup and verification are incomplete. A published pack price is not a statement that checkout is enabled.

Essential authentication cookies keep you signed in. Document translation uses the clipshrink_document_session cookie to identify the browser that owns a task; it lasts up to seven days and is HTTP-only, with secure transport in production. Turnstile may use its own verification mechanisms.

Some tools store preferences or locally edited translation text in your browser. Clearing browser storage can remove those local edits or settings, and clearing session cookies may prevent access to an anonymous translation task. Neither action necessarily deletes the corresponding cloud record.

You can manage cookies and site storage in your browser. Blocking essential storage can prevent sign-in, verification, task access, or other service features from working.

Document and image translation tasks expire 24 hours after creation. The worker is configured to remove expired source and output objects and clear stored text blocks and identifying task fields. Expiry and physical deletion are not the same event: cleanup is best-effort and can be delayed by service interruptions. The source is also scheduled for deletion after processing finishes. The worker removes old daily-quota records after approximately eight days.

Transcription source media is normally deleted from R2 after the worker saves a completed or failed task. Interrupted uploads, tasks that have not started, and failed cleanup or database updates may leave source files for operator recovery. Transcripts, subtitle cues, translations, and task history remain in the account database until a deletion request is handled; there is no automatic 24-hour deletion rule for those records.

Account and credit records are retained while needed to operate the account. Payment, refund, dispute, and security records may need to be retained for reconciliation, fraud prevention, or applicable obligations. This service does not promise one fixed retention period for every database record, support message, hosting log, or provider backup.

Email support to request deletion of an account, transcript, task, or remaining uploaded file. We may ask you to verify ownership and identify the relevant record. We will explain any information that cannot yet be removed and why; deleting local browser data alone does not submit a cloud deletion request.

The service uses Vercel for the website, Railway for processing workers, Supabase for authentication and databases, Cloudflare R2 for private file storage, Cloudflare Turnstile for verification, and Groq or the configured xAI transcription provider for AI processing. Creem handles payments when checkout is enabled. Information is disclosed to these services as needed for their role; their own policies govern information they handle independently.

Provider infrastructure may process information outside your country. Private storage, owner-scoped task checks, limited-life download access, and server-side credentials are used to restrict access, but no internet service can guarantee absolute security or immediate deletion from every provider backup. Avoid submitting confidential, regulated, or highly sensitive material unless you have assessed that this workflow is suitable.

Keep your originals and save the results you need before task expiry. You can use local tools without uploading a device file for processing, choose not to use cloud AI tools, manage browser storage, and contact support about access, correction, or deletion. Your rights under applicable law remain unaffected by this policy.

Changes to this policy will be published here with a revised effective date. For questions about the information used for a particular tool or purchase, contact support before submitting sensitive content or completing checkout.