Files, History and your data

Transcribe an audio or video file#

1. Open File transcription

   Tap Transcribe a file on the Record screen, open Settings and tap File transcription under Tools, or use the Transcribe file app shortcut. You can also share audio or video files from another app to TypeWhisper.

2. Choose the options

   Set the Language (Same as dictation, Automatic or a specific language) and the Profile, or keep None (plain transcription). Turn on Keep audio to play the file back from History. The choices apply to the files you add next.

3. Add files

   Tap Add audio or video files and pick one or more files. TypeWhisper takes any audio or video file whose sound Android can decode, up to 2 GB per file.

4. Open the result

   When a file shows Done · tap to open, tap it to open the transcript in History, or use its copy button.

• Files use the engine from Settings → Speech recognition, unless the chosen profile sets its own.
• Files are transcribed one after another, in short pieces cut at pauses. Long files continue in the background with a Transcribing files notification, and an interrupted file resumes where it stopped.
• Pause holds the queue. Transcribe all continues paused and failed files. Clear finished removes done files from the list; their transcripts stay in History.
• If the app is not open when the queue finishes, a File transcription finished or File transcription failed notification appears.
• If a profile could not be applied, for example because it needs Premium or the writing service failed, the plain transcript is saved and the row says why.
• Detect speakers is a Premium option. It keeps the audio and shows the transcript by speaker. The first time it downloads about 30 MB.

Watch folders for new recordings#

1. Add a folder

   In File transcription, tap Add folder under Watched folders and choose a folder in Android's folder picker. If the recorder app from Samsung, Google, Honor or Huawei is installed, a Watch recordings from … row opens the picker right in its recordings folder.

2. Decide about existing files

   Choose Only new files, or Transcribe them too to include the audio and video files already in the folder.

3. Set the folder's options

   Tap the folder to give it its own Profile. Turn on Save transcript next to the file and choose under Save as: Text (.txt), Subtitles (.srt) or Text and subtitles (.txt and .srt). Existing files are never overwritten.

• New audio and video files, also in subfolders up to three levels deep, are transcribed with the language, speaker and audio choices above.
• TypeWhisper checks about every 15 minutes and whenever you open the app. Files that are still being recorded wait until they are finished.
• Below the folder list, turn on Only while charging or Cloud services only on Wi-Fi to hold new files back. Both are off by default and apply to all folders. With on-device speech recognition, files do not wait for Wi-Fi. Files already in the queue are still transcribed.
• Each folder row shows when it was last checked or why files are waiting. Use its switch to pause watching, or the remove button to stop watching it.
• Results go to History, not to the Inbox.

Find transcriptions in History#

• Search history finds words in the text, in the original before corrections, in file names and in profile names.
• Filter chips show Inbox while quick dictations wait, All, and one chip for each source you have used: App, Keyboard, Bubble, File and Quick dictation.
• Each row shows the date, the time and the engine. An icon shows where the entry came from.
• The export button at the top shares all entries as Markdown (.md), Text (.txt) or JSON (.json).
• Touch and hold an entry to select it. Tap more entries or All, then mark them as done, export or delete them.
• Swipe left to delete an entry. Swipe right to mark it as done or to put it back with Move to Inbox.

Review an entry#

• The line above the text shows the date, the engine, the length and the profile.
• If TypeWhisper changed the engine's text, for example with the dictionary or a profile, the original appears below as Before corrections: ….
• Copy and Share sit at the top. Emails written by a profile get an Email draft button, calendar drafts get Add to calendar.
• The ⋮ menu has Edit text, Add correction, Move to Inbox and Delete. After you edit the text, TypeWhisper offers to add the changed words as corrections.
• Export subtitles shares the transcript as SubRip (.srt) or WebVTT (.vtt). It is available for file transcriptions and for dictations whose recording was kept.
• For text written by a profile, Report AI text prepares an email to the TypeWhisper team with the profile, your dictation and the written text. You see it in your mail app before sending.

Playback, speakers and transcribing again#

Dictations keep their audio only if Keep recordings in Settings → General was set to 7 days, 30 days, 90 days or Keep forever when you recorded. The default, Delete after transcription, keeps no audio. Audio from files is kept when Keep audio or Detect speakers was on, and stays until you delete the entry.

• With audio, the entry shows a player: play and pause, 10 seconds back or forward, and a speed button that switches between 1×, 1.25×, 1.5×, 2× and 0.75×.
• Transcribe again in the ⋮ menu runs the kept audio through an engine you choose. It replaces the text and the engine; the profile is not applied again, and detected speakers are reset.
• Detect speakers in the ⋮ menu needs Premium. Choose the Number of speakers or leave it on Automatic, then tap Detect. Detection runs on the phone.
• The Speakers tab shows who spoke when; the Text tab shows the plain transcript. Tap a turn to play it, or touch and hold it to assign it to another speaker.
• Name speakers saves names with the recording. Recognize this voice in future recordings creates a voice profile that stays on this phone. Save a voice only with the person's consent.

Recover failed or interrupted recordings#

TypeWhisper writes each recording to the phone while you speak. If a transcription fails, for example without a network connection, or the app is closed during a dictation, the recording is kept for 24 hours. Recordings without speech and dictations from password and incognito fields are not kept.

1. Open Recoverable recordings

   When there is something to recover, the row sits at the top of History. It says how many failed dictations can be transcribed again, or reads The last dictations, kept for 24 hours.

2. Try again

   Tap Again next to a recording and choose an engine. The new text is copied to the clipboard and opens in History.

• Unless Settings → General → Keep recordings is set to Delete after transcription, the last three successful dictations are kept here too, so you can try them with another engine.
• Recovered quick dictations go back to the Inbox.
• The delete button removes a recording right away.

Back up, export or delete your data#

Open Settings and tap Data and backup under Help.

• Back up settings saves settings, profiles, snippets and the dictionary as a JSON file. API keys are never included.
• Restore a backup reads that JSON file. It adds missing profiles, snippets and dictionary entries, keeps existing ones and applies the saved settings. Enter your API keys again afterwards.
• Export all data saves the backup plus History, speaker transcripts, kept recordings and statistics as a ZIP file.
• Delete all data removes History, recordings, dictionary, snippets, profiles, voice profiles, watched folders, statistics, settings and API keys from this phone, signs you out of your TypeWhisper account here and restarts the app. Downloaded models stay. This cannot be undone.
• Voice profiles are biometric data and stay out of every export and backup.
• In Settings under Tools, Statistics shows words, streaks, words per minute, time saved, sources, engines and usage by time of day for a week, a month or all time. It stays on this phone, and Clear statistics leaves History and recordings untouched.