VoiceVibe

Tap the logo or Tap to Speak and allow your microphone. Speak naturally, then tap Stop. Your words appear in the transcript after processing. Correct anything there and tap Record more to continue.

Whisper transcribes on your device. On first use, the browser may download substantial model files (roughly 100 MB, varying by model and browser). Later visits can reuse cached files. Internet access is required for initial setup unless the model assets were installed on the same site. Keep the page open while it prepares or transcribes.

Each clip can be up to five minutes. Longer clips take longer to process; speed varies by device. If you leave the tab while recording, the clip stops and begins processing. Tap Cancel to discard the current clip; your existing transcript stays.

Open Import audio or try a sample to choose an audio file up to five minutes or 40 MB. Transcription prepares automatically and adds words to your draft. The sample is a short English speech clip.

Play back the latest clip using the audio control. Save recording downloads that clip. If processing fails, the audio stays available while this page is open: tap Retry transcription or save it before leaving.

Editing your draft during recording or processing keeps your corrections. New words append at the end. New Transcript and Delete cancel pending work, so old results cannot enter a new draft.

The text box is your working draft: correct words, add paragraphs or paste notes. The meter shows the word count and an approximate speaking time at 140 words per minute.

New clips are blocked at 15,000 words. The full text of a completed clip is kept, including overflow. Save Markdown, then start a new transcript to record more.

Clean Spacing removes extra spaces and blank lines without rewriting your words or punctuation. Select Undo within ten seconds to restore the exact text. A later edit or a new recording ends that undo.

Content Type labels your notes and helps name downloaded files. Start with Ideas & Notes, Copywriting, Build Notes or Audiobook.

Choose + Add content type… to create your own. You can keep up to ten custom types with unique names of 1–40 characters. Use Manage to rename or delete them. Deleting the selected custom type switches the draft to Ideas & Notes.

Prompt Type chooses the instructions included with your Markdown. Read the short description, then select Preview Prompt to see the complete instructions. The preview uses a placeholder for your transcript.

Copy Prompt copies those instructions only. To include your actual words, use Copy MD or Save MD and paste the result into your chosen AI tool. VoiceVibe does not send the transcript to an AI service for you.

VoiceVibe logo

Copies only your current transcript. The logo stays highlighted after a successful copy until the text changes or you leave the page.

Copy Prompt

Copies the previewed AI instructions with a transcript placeholder.

Copy MD

Copies the complete Markdown package: transcript, session details, dates and selected AI prompt.

Save MD

Starts a Markdown file download with the same complete package. The filename uses the Content Type, date and running save number.

If copying is blocked, use Save MD or select and copy text manually. Check your browser’s Downloads to confirm the file was saved.

New Transcript starts a new draft after confirmation. Copy or save first if you want to keep the old one. Your custom types and remembered selections stay available; New does not increase the save number.

Delete clears the transcript and its browser draft after confirmation. Select Undo within ten seconds to restore its text, session, selections and dates. Undo is temporary: it ends after the timer, another edit, New, reload or leaving the page.

Your nonempty draft is saved automatically in this browser when storage is available. Reopening the same page in the same browser/profile restores the text and its selections. Clearing all text clears the saved draft too.

Created stays with the session. Last Updated changes when you edit text or selections. Restoring a draft does not reset its Created time.

Save or copy your work before updating or moving the app. Another browser, profile or file location may have a separate draft. If a storage warning appears, Save MD before closing the page.

Audio is processed on your device with Whisper and is not sent to a transcription service. The app downloads its software and English model. There is no app account or cloud draft sync. Drafts remain in this browser until you copy, save or clear them.

Use the HTTPS website and allow microphone access. If recording fails, check browser permissions, return to this tab and tap Record again. If transcription fails, check your connection, close heavy tabs and use Retry transcription. Typing, copying and Markdown export remain available.

Recognition can make mistakes. Review names, numbers and punctuation before using your transcript. Accuracy improvement over the previous engine has not yet been measured with matched personal recordings.

Save Markdown downloads your transcript plus the selected prompt to a .md file. Copy Markdown sends the same content to the clipboard; check browser clipboard permissions if needed. Use Save recording after recording if you also need the original audio.

Download VoiceVibe: Download complete release package (ZIP). This package includes the application, embedded images, Whisper browser runtime, WebAssembly, source modules, licenses and a model-setup helper. Whisper model weights are not embedded and must be installed with the helper for a fully local deployment.

For a working installed app, extract the complete ZIP into your web server directory rather than opening the HTML in isolation. Microphone access requires HTTPS or localhost. The standalone HTML can render without a server, but its transcription runtime may require online resources.