ScribeKey privacy policy

Effective 26 October 2026

If the app behaves differently from this policy, email support@scribekey.app.

About ScribeKey

ScribeKey is an Android app. It does not run a ScribeKey account service or keep a server-side user database.

What stays on your phone

All of the following is stored in ScribeKey's private app storage, which no other app can read:

• Transcripts. Capture sessions and ordinary dictations are saved locally. Your edited text is stored separately from the original transcript. Continuing an existing session or dictating an editing instruction does not create a separate history entry.
• Audio recordings. Captured dictations and retained imported recordings are saved locally for playback and re-transcription. Saving dictation audio is on by default. Turn it off with Keep audio, in Settings under Privacy & security. A recording still in progress or waiting to be processed is kept until it becomes a session.
• Your dictionary, vocabulary, rewriting prompts and voice snippets. Vocabulary can also be included in an optional cloud rewrite request, as described below.
• The app you dictated into. Its name and package are recorded on each entry so your history shows where the text went.
• Your statistics. Words dictated, time saved and speaking speed are calculated from local history.
• Your settings, including any API keys, which are encrypted with a key held in the Android keystore. Speaker names you add for Capture are also stored in settings and included in an optional encrypted backup.
• Meetings, speakers and kept voices. When you ask ScribeKey to find the speakers in a recording, it works them out on your phone and writes their names into the note; the sound of each voice is not saved with the session. After you agree to remember named speakers, naming someone remembers a voice profile when enough usable audio is available. A voice profile is a set of numbers describing how they sound, which is biometric information. It is made only after that consent, only for someone you named, and only on your phone. An existing person's name requires confirmation before their voice profile gains a sample. Paused voices stay paused. Voice profiles are kept in their own file that Android backup, phone-to-phone transfer, ScribeKey backups and exports never include, and they are never sent to any server. Ask the person before you keep their voice. If you tap Find speakers · Cloud and confirm, or choose it for every meeting in Settings under Meetings, the recording is sent with your own OpenAI or Microsoft Foundry key to that service, which returns who spoke when; voice profiles are still made and kept only on your phone and are never sent.

Nothing here is copied to Google's servers unless you turn on Android backup in Settings under Backup. It is off until you do. When it is on, Android backs up your sessions, dictionary and skills, but not your audio, settings or API keys, end-to-end encrypted with your phone's screen lock, so Google cannot read them; without a screen lock nothing is backed up. When you set up a new phone from your old one, Android moves your transcripts and saved audio directly between the two phones whether or not Android backup is on. Your settings, API keys and voice models stay behind. Set them up again on the new phone.

What can leave your phone, and only if you turn it on

On device only, in Settings under Privacy & security, pauses cloud speech, cloud skills and cloud speaker finding until you turn it off. It does not pause model downloads, links you share, skill lists, note sync, backups you save to a folder or Android backup.

Model downloads

Speech recognition runs on a model file that has to be downloaded once. ScribeKey downloads it from huggingface.co when you choose a model during setup or later in Settings. The model that tells speakers apart in meetings is downloaded from Hugging Face and GitHub.

The download is an ordinary file request. ScribeKey sends no identifier of any kind with it: no account, no device ID, no usage information. When you open the app, it checks GitHub for updates to the list of available models. If you add a custom model, the text you search for and the repository ID you enter are sent to the Hugging Face Hub. Browsing rewriting skills fetches skill lists from GitHub, or from a skills marketplace whose address you add. All of these requests are anonymous and require no sign-in. As with any web request, the servers see your IP address and the time of the request.

Links you share

When you share a link to a podcast episode or a recording with ScribeKey, it downloads that page and the recording it names, the same way a browser would. For a Spotify episode, which only offers a preview, it sends the episode's title to Apple's public podcast directory to find the show's own file. These requests carry no account or identifier, and the downloaded copy is deleted once the recording is read.

Cloud transcription is off by default

ScribeKey can send audio to an OpenAI-compatible transcription service or Microsoft Foundry instead of transcribing on the device. This is off unless you switch it on and enter your own API key.

• You choose the endpoint. There are presets for OpenAI, Groq, Mistral and Microsoft Foundry, and you can type any other compatible URL. A Foundry streaming model receives the audio while you speak.
• What is sent. The request body contains the audio being transcribed, your selected model and, if you set one, the spoken language. This includes imported audio and saved recordings you retry. ScribeKey adds no app name, device ID or history; your API key authenticates the request.
• Where it goes. Audio goes directly to your configured endpoint. ScribeKey does not relay audio through its own server.
• Your API key is stored encrypted on the device and is sent only to the endpoint you configured, to authenticate that request. Remove it in Settings under Cloud connections.

Once you send audio to a third-party service, that service's own privacy policy governs what it does with it.

Cloud skills are off by default

Skills rewrite text, separately from speech recognition and ordinary cross-app cleanup. Set up a provider and review cloud access before using it. A rewrite runs when you choose one. It also runs without a tap if you turn on a skill for a particular app, so dictation into that app is rewritten in the cloud, or turn on a skill to run after meetings. Both are off until you turn them on, and you can turn them off in Settings. A rewrite sends the text being edited, your instructions, and vocabulary supplied to the rewrite to that provider. Where app context is supplied, the app name and package can be included too. The configured provider's own policy governs that processing.

Choosing on-device speech does not make a later cloud rewrite local.

The rewriting model never receives the contents of your screen. See the next section.

Note sync is off by default

Backup & sync, in Settings, can copy your sessions as readable notes to places you choose. Nothing is copied until you set one up.

• Notes folder writes each session as a Markdown or text file, and its recording if you choose, to a folder you pick. That folder can belong to a storage app that uploads it to that app's cloud.
• WebDAV server sends each session as a note, and its recording if you choose, over HTTPS to a server address you enter, such as Nextcloud or a home NAS. Sync runs in the background every few hours. The server password is stored encrypted on the device and sent only to that server.

These copies are not encrypted by ScribeKey. Deleting a session in the app does not delete a copy already written to your folder or server; remove it there. The server or storage app's own policy governs what it does with your notes.

Optional Refine cleanup on device

Refine downloads the ScribeKey Fast model and cleans up English text locally on supported devices. Transcript text is not sent to the model host. Basic cleanup is the fallback when the local model cannot complete safely. You can remove the cleanup model in Settings.

The accessibility service

ScribeKey uses Android's accessibility service to type your words into supported editable fields in the app you are using. Android shows this permission with a broad warning, because the same permission can, in principle, be used to read the whole screen. Before the first handoff to Android's Accessibility settings, ScribeKey shows a prominent disclosure with Continue to Accessibility settings and Not now. ScribeKey records the disclosure version and acknowledgement time only after you continue. You can turn this access off at any time.

What it does:

• Notices when a text field has focus, so the dictation button knows when to appear. To decide, it reads the field's hint, ID and input type, not its text.
• Reads the focused field's text and selection, only as needed so dictated text replaces a selection or lands at the current cursor.
• Writes the transcribed text into the field.
• Looks up the name of the app that has focus, so your history can show where each dictation went.
• After a dictation lands, follows that one field for up to two minutes, so a word you correct there can be suggested in your Dictionary. It stops when you move to another field, send or clear the text, or dictate again, and only does this while Learn from my edits is on.
• Checks whether a field is a password field, before anything else happens to what you said. Nothing spoken at a password field is typed, rewritten, sent to a cloud service or saved. The recording is deleted and your history shows only that a dictation was refused.

What it does not do:

• It does not read password-field contents or unrelated screen content.
• It does not save or send the focused field's text or selection. The name and package name of the app you dictate into are saved locally with the session. From a correction to dictated text it keeps only the changed words, about three words either side and the app's name, on your device, as a suggestion you add or dismiss. Suggestions are not included in backups.
• It does not read your clipboard.
• It does not log keystrokes or take screenshots, and beyond the field it just dictated into, it does not follow what you do in other apps.
• The accessibility service does not send anything anywhere. It collects only the name and package name of the focused app and the correction suggestions described above, stored on your device. Audio and text leave the device only through the cloud and sync features described above, when you turn them on.

Crash reports and analytics

ScribeKey includes no crash reporting, analytics or performance monitoring. This applies to every build, including internal builds labelled ScribeKey QA.

Android tester recruitment

If you join the Android testing cohort, ScribeKey collects the Google account email address you use on Google Play to send you testing invitations and closed-track installation instructions. This information is used solely for testing coordination through Google Groups and Google Play, which process it under Google's terms. It is not shared otherwise, and is retained only during the testing cycle. You can leave the testing group or request removal at any time.

Google Play purchases

If you buy ScribeKey, Google Play handles the payment. The app checks Google Play's signature on the purchase on this device; nothing about the purchase is sent to ScribeKey, and ScribeKey runs no server for it. The app stores only when your trial started, whether ScribeKey is unlocked and when Google Play last confirmed it. Once Google Play has confirmed a purchase, ScribeKey stays unlocked while you are offline; only an answer from Google Play, such as a refund, locks it again. Google Play processes purchase information under Google's own terms and privacy policy.

Permissions and why they exist

• Microphone: to hear what you dictate. Recording only runs while a dictation is in progress, with a notification showing.
• Accessibility service: to type into other apps. Described above.
• Display over other apps: for the floating dictation button.
• Notifications: so you can stop a recording from the notification shade, and to show download progress.
• Internet: for model downloads, and for the cloud and note sync features if you turn them on. Local transcription works without a connection once its model is downloaded.
• Fingerprint or screen lock: optional, to lock ScribeKey. Android checks it; ScribeKey stores nothing about it.
• Background work: to finish downloads, processing, backups and note sync after you leave the app.
• Vibration: for touch feedback.
• Ignore battery optimisation: optional, so long dictations are not killed mid-sentence.

How long things are kept, and how to remove them

• Transcripts stay until you delete them. Delete one from your history, or select several and delete them together.
• Audio files stay on your phone until you delete the recording they belong to. You can instead set a retention period in Settings under Privacy & security. ScribeKey then sweeps once a day and deletes audio older than it. Android runs the sweep, so a file may last a few hours past the period. With a period set, Delete expired audio now clears expired audio immediately.
• Your API keys are removed in Settings under Cloud connections, one at a time or all at once.
• Voice profiles stay until you forget them. Meetings, in Settings, lists everyone whose voice is kept, with a way to forget one person or everyone. Forgetting a voice deletes it from your phone; names already written into your notes stay.
• Uninstalling ScribeKey removes its private app data and downloaded models. On Android 10 and later, Android may offer to keep the app data instead; if you keep it, reinstalling ScribeKey picks it up again. Copies you exported or shared, backups you saved elsewhere, notes synced to a folder or server, and data already sent to a cloud provider are not removed by uninstalling the app.

There is no single "delete all my data" button in this version. You can delete sessions in the app and remove exported copies from wherever you saved them.

Exporting and backing up your data

You can export transcripts as plain text, Markdown, JSON, CSV, or SubRip and WebVTT subtitles, share a session's recording, and you can create a backup protected with a password you choose, containing your transcripts, dictionary, prompts and, if you choose, your audio. Exports go through Android's normal share sheet, so you decide where they end up. You can also have ScribeKey save a protected copy each day to a folder you pick, on your phone or in any storage app that offers folders. API keys are deliberately left out of backups. Keep the backup password somewhere safe. ScribeKey cannot recover it for you.

Children

ScribeKey is not directed at children under 13. Local data is managed in the app.

Changes to this policy

If the app's behaviour changes, this page changes with it, and the effective date at the top moves. Material changes will be called out in the app's release notes.

Contact

Email support@scribekey.app with questions, corrections, or anything on this page that does not match what the app does.