Your music. Your data.

What is stored

Notavue stores submitted YouTube links or audio uploads, titles, source hashes, generated versions, processing times and failure reasons. Profiles, identity links and score ownership are stored in Replit PostgreSQL. Audio, scores and analysis evidence are stored in private Replit App Storage.

Sign-in does not use passwords. Available methods reflect current provider configuration. When email codes are enabled, the email provider processes recipient addresses and verification messages. Google handles sign-in verification when enabled. Source recordings are not sent to the sign-in email provider.

Marketplace data and payments

Hima Inc operates Notavue. After you choose to list a score and it passes review, its title, composer, editor, price and cropped preview become public and may be read by search engines and AI services. Source recordings and complete private files do not become public.

Stripe handles Checkout, identity and bank verification. Notavue stores order identifiers, payment status, actual fees, revenue shares and payout reconciliation records. Our forms do not collect complete card numbers, bank passwords or identity-document images. Purchased editions are delivered only to buyers who remain authorized.

Only your explicit permission allows title and composer searches in public stores; otherwise comparison uses anonymous score features and adjacent categories. The optional OpenAI pricing model receives public price facts and comparison bounds, not your score, audio or account data.

Deleting a creator account stops new sales. Fixed editions already lawfully delivered to buyers and transaction records needed for settlement, refunds, disputes or legal duties are retained as necessary, separately from private source recordings, practice data and public listing access.

Who can access it

Source recordings and full downloads remain private. When creating a link, you may choose to include synthesized piano playback of the preview. Personal fingerings, notes and practice positions are not shared.

Authentication records are cleaned periodically: verification challenges remain for 7 days after expiry, and expired or revoked sessions for 30 days. Your account and scores are retained.

Works are private by default. Before sign-in, access uses a secure credential in the current browser; verified accounts use account ownership. Signing in grants access only to works you are entitled to open. Clearing browser data may remove access to unclaimed works.

People with a share link can read the title and preview only after you create that link. Links last 30 days and can be disabled; source audio and full downloads are excluded. Search and AI crawlers can access public product pages. Private APIs, accounts and scores are not publicly indexed.

Manage and export

After signing in, save your practice position, favorites, notes and fingering annotations, and access them from other signed-in devices. Personal annotations are stored separately from the original transcription and are not evidence of AI accuracy or publication approval.

Download profile, source and score records as JSON from Profile. Audio and score files are excluded; available score files can be downloaded from the score page. Only data owned by the current credential or verified account is included.

Archiving can be undone; permanent deletion cannot. Your profile shows the deletion option once enabled. Download any files you wish to keep and wait for current tasks to finish. Confirmation disables all sessions and shares. The system erases private audio, scores, analysis records and the original identity; the receipt shows actual progress.

Daily database backups follow a 31-day retention policy. Security receipts keep purpose-limited hashes and anonymous identifiers so restoring a backup cannot revive deleted data. The mailbox-equivalence hash used against repeated referral rewards remains for 7 days. Credits already earned by other users are preserved.

Record audio

Recordings stay in the current tab until you choose Generate score. Stopping, cancelling, changing sources or leaving the page closes the microphone; recording stops when the tab moves to the background. Unsubmitted recordings are not saved across devices.

Contact support

Support emails are sent, received and stored through Resend to handle usage and payment questions.

Service records and usage rights

Job status, performance and errors are recorded to improve reliability and limit abuse. Operational health summaries contain no titles, source URLs, account or job identifiers. Original processing records remain private.

Use recordings and compositions you have permission to process. Automatic transcription does not license a composition, recording or arrangement. Confirm the necessary rights and musical quality before public performance or commercial use.

Optional usage measurement

Measurement may aggregate the first observed page category in a tab and subsequent signup clicks, sign-ins, playback, sharing and note saves. These are consented browser events, not independent people, verified backend registrations or payments.

Off by default. If you consent, aggregate page-category and preview/full-piece workflow counts help improve the service. Measurement does not store titles, links, audio, email, IP addresses or detailed browsing history, and uses no third-party analytics. Connections are still processed by the website host.

You can turn this off anytime. A random ID lasts for this tab session; the server temporarily holds deduplication values for the current day. Replit stores aggregate counts, not the identifier. Turning it off stops new events; already anonymized aggregate counts cannot be removed individually. Do Not Track or Global Privacy Control prevents collection.