The short version
MP3 to Transcript processes the audio you choose so it can return a transcript. The application does not save the raw upload to its database, publish it, or create a public transcript link. Audio is uploaded over HTTPS to private temporary object storage, sent to the configured transcription provider, and deleted when processing completes, fails, or is cancelled. A maximum 24-hour lifecycle is the cleanup backstop. For signed-in users, the resulting transcript text and timestamps are saved privately for the retention period shown by their plan.
Only upload audio that you own or have permission to process.
Audio processing
- Your browser sends the selected audio file to our server after you choose Create transcript.
- The server validates the file type and size, writes resumable upload parts to private temporary object storage, and starts a recoverable transcription job.
- Long MP3 files are split only at MPEG frame boundaries into provider-safe temporary chunks. Upload parts are deleted after preparation; provider chunks are deleted when the job reaches a terminal state.
- The provider returns the transcript, speaker segments, and timing data when supported by the selected model.
- Our application does not write raw audio to its database or expose it through a public URL. Temporary source parts and processing chunks are deleted as soon as their stage ends, with a maximum 24-hour storage lifecycle as a failure backstop.
- For signed-in users, we store the resulting transcript text, speaker and timestamp segments, filename, language, duration, usage amount, processing state, model name, and retention deadline.
- Copy and download actions happen in your browser. You decide where the result goes after that.
Our default provider is fal's Whisper API. Source-linked summaries and translations use fal's OpenRouter endpoint. Provider handling is also subject to fal's and the selected model provider's terms and privacy documentation.
Technical and usage data
Hosting, security, and analytics services may receive routine request data such as IP address, browser information, timestamps, error details, and page events. We use that information to keep the service reliable, prevent abuse, and understand aggregate product usage.
We do not intentionally put raw audio or full transcripts into analytics events or application logs.
Accounts and payments
Account providers may share basic account details such as name and email address. Stripe returns customer, checkout, subscription, invoice, and payment status references so we can determine the plan and billing period. We do not receive or store full card numbers.
Retention and deletion
Raw audio is not retained in the application database or transcript history. Temporary upload parts are deleted after preparation, processing chunks are deleted at terminal status, and all temporary audio has a maximum 24-hour object storage lifecycle. Saved transcript text and timestamps are automatically deleted after the plan retention period: 7 days on Free, 30 days on Starter, 90 days on Pro, and 365 days on Studio. A signed-in user can delete a transcript earlier from the private transcript library.
Usage and payment metadata may be retained longer where needed for account operation, security, fraud prevention, support, financial records, or legal obligations. The configured transcription, hosting, security, analytics, and payment providers control retention within their own systems and policies.
Security
No internet service can promise absolute security. We use HTTPS, server-side API credentials, file validation, and reasonable technical safeguards. You should avoid uploading highly sensitive recordings unless you have assessed whether the service and configured provider are appropriate for that material.
Contact
Privacy questions can be sent to hello@mp3totranscript.com.