Transfix is a web-based tool designed to transcribe speech recordings directly within your browser. It's novel user interface allows you to quickly correct any mistakes by clicking on the alternative words that are presented.

Transfix supports whatever audio and video files your browser can play. Most modern browsers support .wav, .mp3, .mp4, .mov, .ogg, .webm, and .flac.

The accuracy of transcriptions depends on the quality of the audio recording and clarity of speech. TransFix uses a simple, lightweight speech recognition system that can run locally, so some manual correction will be necessary. However, this can be done quickly thanks to the novel correction interface.

There are four ways to edit the transcript:

• Select
  If the correct word is listed below, click on it to switch to that word.
• Skip
  Click the X button above the word to skip it.
• Insert
  Click the arrow between words to insert a new word. Type the word and press return.
• Replace
  To replace an existing word with a new one, select it then click on it again. Type the new word and press return.

We currently only support US English, but hope to add support for other languages in the future.

The length is only limited by the processing speed and memory availability of your machine. Low-spec computers and mobiles may grind to a halt as they are unable to transcribe quickly enough, which creates a backlog.

When the transcription is complete, you can click the copy button to copy the corrected transcript to your clipboard for use in other applications.

No - Transfix processes your recordings directly in your browser, so your audio files and transcriptions are not uploaded to any server. All processing is done locally in your browser to maintain your privacy.

Yes, an internet connection is required to access the Transfix web tool. Once the page has completed its initial loading, you can disconnect from the internet. We hope to add full offline support in the future.

Although Transfix is designed to work on desktops, it should also run on high-spec mobile devices.

It is completely free to use.

We do not collect or store any personal data, or any data from your audio files or transcriptions. We do collect anonymous website traffic data, for understanding visitor numbers and behaviour. We do this using a privacy-focused, open-source analytics solution called Umami. Users cannot be identified and are never tracked across websites. The data is stored on own servers located within the EU. We have total ownership over this data and do not share it with anyone.

To make this, we used Svelte, Vite, daisyUI, Tailwind CSS, PostCSS, core-js, cytoscape.js, and theme-change (all MIT); Vosk, vosk-browser, and TypeScript (all Apache 2.0); and elkjs (Eclipse Public Licence 2.0).

Please click the feedback button on the navigation bar to share your bug reports and ideas.