Privacy and your data

What stays on your device

ReelTranscript has no cloud backup or cross-device sync. Imported audio is saved in IndexedDB in your browser. Lesson details, transcripts, saved vocabulary, preferences and personal API keys are stored locally using IndexedDB or localStorage.

Local time-based splitting and playback do not upload your audio. Anyone with access to your browser profile may be able to access its saved data, including keys. Use a trusted device.

When you choose to visit your account from an import, the selected file and split settings are saved as a local draft for up to one hour so you can return after signing in. Expired drafts are removed when next accessed; you can also discard the selected file. Drafts are not uploaded.

Optional Google account and usage limits

Where sign-in is configured, Supabase Auth handles Google authorization and stores your account identity, email and basic profile. We do not request access to Gmail messages. The browser stores access and refresh tokens in localStorage to keep you signed in. Sign out on shared devices; signing out does not delete local lessons.

The online allowance service uses an HttpOnly signed anonymous trial cookie, a keyed hash of your network address, account ID, request identifiers, durations, times and processing states to limit usage and abuse. A network hash is an abuse-control signal and may group people sharing a connection. These quota records are separate from the broad action events below. No audio, transcripts or provider keys are stored in the quota ledger.

Platform-funded transcription for imported local audio is not connected in this release. The separate home-page Instagram Reel tool uses Cloudflare Workers AI. Video-link importing is a separate optional service and does not require Google sign-in. Your own provider key remains optional and is billed by that provider. Account deletion and quota-record requests can be made through the maintainer; clearing browser data alone does not remove server account records. Records remain until removed by the maintainer, subject to the service provider’s retention and backup policies.

Optional video-link imports

When you submit a YouTube or TikTok link, our extraction service receives that public video URL and requests its media from the platform. A keyed hash of your network address is used for rate limits. The service temporarily stores extracted audio, its title and task status for up to 30 minutes after completion, then removes them on its periodic cleanup. Cancelling in the browser stops waiting; an already queued download may still finish and expire normally. Service restarts discard temporary tasks.

The resulting MP3 is downloaded into this browser and follows the same local storage and optional transcription rules as an uploaded file. Import only videos you own or have permission to use. Hosting and platform providers may retain ordinary request logs under their own policies.

Instagram Reel transcription

When you submit a public Instagram Reel URL on the home page, our server retrieves its audio and sends it to Cloudflare Workers AI to create a transcript. The result stays on the page until you leave or refresh; use Copy text or Download SRT to keep it. This is separate from your local listening library and does not require your own Gemini or Deepgram key.

Import only content you own or have permission to use. Stopping the wait in your browser may not stop server processing. Hosting, Instagram and Cloudflare may retain ordinary request logs under their policies.

Optional transcription

When an online provider key is configured, creating a listening session sends your audio directly to the selected transcription provider, Gemini or Deepgram, including when the chosen section method is By time. Word alignment also sends the original audio. Your key is stored in this browser and sent directly to that provider. Provider charges, retention and data handling follow that provider's policies.

To use time-based splitting without uploading audio, remove your provider key in Settings and save before importing. Choosing the explicit time fallback after a failed transcription does not send another transcription request.

Other connections

Looking up or translating a word may send that word to dictionary or translation services. The website also loads interface resources, fonts and styles from external CDNs and font services. These services receive ordinary network request information such as your IP address.

Basic usage and reliability

The website sends broad action events to its own /api/events endpoint, such as trying a sample, importing audio, completing a split, starting playback or encountering a transcription error. Event payloads contain only action names and the app version: no audio, transcript text, file names, API keys, email addresses or user identifiers.

The hosting platform may keep ordinary request logs, including network information. These events help identify where the tool fails or is difficult to use.

When you contact the maintainer

The feedback link opens your email app to send a message to dihou008@gmail.com. Nothing is sent until you send the email. Your email address and the details you choose to include are used to respond to your message.

Audio, transcripts and API keys are not attached automatically. Do not include keys or private recordings in feedback. You can review and remove any prefilled details before sending.

Keeping or removing your data

Keep your original audio files outside the browser. Clearing site data, using private browsing or browser storage cleanup can remove your library permanently.

• To remove a provider key, clear its field in Settings and save.
• To remove all local ReelTranscript data, use your browser's site settings to clear storage for this website. This removes lessons, audio, vocabulary, preferences and saved keys together.
• Before clearing storage or changing devices, keep copies of any original recordings you need. There is no server copy to restore.

Questions about data? Contact the maintainer.