Voice notes
Capture ideas quickly
Speak a thought, reminder, or rough draft instead of typing it out. Record another note to append it to the same text.
Private · English-only · Free
Turn speech into
text, effortlessly.
Record a quick thought or upload an audio file. Lexiconic turns your English speech into text right in your browser — then lets you take it with you.
No account
Audio stays on device
Free to use
Up to 10 min per clip
Ready when you are
01 CAPTURE AUDIO
Start with your voice.
Record a note, or bring an audio file you already have.
Tap to record · tap again to finish
Maximum recording time 10:00
OR UPLOAD A FILE
English speech · up to 10 minutes per clip · compatible formats depend on your browser.
02 YOUR TEXT
The final transcript.
Finished transcripts appear here. Edit them, copy them, or download a plain text file.
Preparing your audio…
The first use downloads the local English speech model. This can take a moment.
Local by design. The English AI model downloads only when you first use transcription (about 175 MB of model weights). Your audio is processed on your device and is not uploaded to a transcription server.
Made for quick, useful text.
A lightweight way to turn spoken English into a draft you can actually use.
Writing
Get a first draft
Dictate an email, outline, or speech and edit the final text in place before copying or downloading it.
Audio files
Turn recordings into text
Upload an MP3, WebM, WAV, or another browser-compatible audio file and receive an editable transcript.
Privacy
Keep audio on your device
Transcription runs locally in the browser after the model is loaded. No account or transcription upload is required.
Free English audio to text, right in your browser
Lexiconic is a simple speech-to-text workspace for short recordings, voice notes, and audio files you want to turn into editable writing.
Convert MP3, WebM, and voice recordings into text
Use Lexiconic when typing is slower than speaking. Record an idea with your microphone or upload an existing MP3, WAV, M4A, OGG, FLAC, or WebM file. When your browser can decode the audio format, the English speech model transcribes it and places the finished text in an editable document. You can correct names, punctuation, and wording before copying the result into an email, a document, meeting notes, or a draft.
Each clip can be up to 10 minutes long. If you have several short recordings, process them one at a time and Lexiconic will append each completed transcript to the same document. When you are finished, download the combined text as a simple TXT file.
Private, local-first speech-to-text
Lexiconic runs its English speech recognition in your browser instead of sending the recording to a transcription service. The AI model is downloaded from the model host the first time you use transcription and may be cached by your browser for later visits. Your audio is decoded and processed on your device, so the recording itself does not need to be uploaded to a transcription server.
Because the model runs on your device, processing speed depends on your computer, browser, memory, and the length and clarity of the recording. For best results, use clear English speech, reduce background noise, and review the transcript before relying on it for names, figures, or important details.
How to transcribe audio to text with Lexiconic
The workflow is designed to be quick: capture your speech, let local AI process the clip, and take away the text.
Record or upload audio
Press the microphone to start a voice note and press it again to finish, or choose a compatible audio file from your device. Microphone access requires a supported browser and permission.
Let the English model transcribe it
On first use, Lexiconic downloads the local model. A progress indicator shows model loading and transcription progress. The finished transcript appears after all audio segments have been processed.
Edit, copy, or download TXT
Read through the transcript, make corrections in the text area, copy it to your clipboard, or download a .txt file. Start another short recording to append more text to the same document.
Tips for a clearer English transcript
Speak naturally in a quiet space
Place the microphone close enough to capture your voice clearly, speak at a comfortable pace, and avoid music or several people talking at once. Recordings with echo, low volume, overlapping speakers, or heavy background noise can be harder for an automatic speech recognition model to understand.
Review the text before using it
Automatic transcription creates a useful draft, not a guaranteed verbatim record. Check proper names, numbers, technical vocabulary, and punctuation, especially when preparing professional notes or a speech. Lexiconic lets you edit the final transcript directly before you copy or save it.
Frequently asked questions
A few details before you start.
Each file or microphone recording can be up to 10 minutes. You can transcribe additional clips one after another and append the finished text to the same document. The browser's available memory still matters.
Lexiconic is optimized for English speech. Other languages may produce unreliable results, so this version focuses on English.
No. The app downloads the model from Hugging Face when first needed, but the audio itself is decoded and transcribed in your browser.
Lexiconic downloads the model automatically on first use (about 175 MB of model weights). The browser may reuse its cache for later visits; cache retention depends on the browser.
Try MP3, WAV, M4A, OGG, FLAC, or WebM. Actual decoding support depends on your browser and the codecs inside the file. If a file is rejected, convert it to a common audio format and try again.
Yes. Upload a compatible MP3 or another supported audio format, wait for the English transcription to finish, then edit, copy, or download the final text as a plain TXT file.
Yes. After one clip finishes, record another or upload a new file. Each completed transcript is appended to the editable text, separated by a blank line.
Yes. There's no subscription or account. Processing speed depends on your device, and the initial model download requires an internet connection.
Not yet. Recording happens live, then Lexiconic transcribes the full clip after you stop. A single progress bar shows processing progress; the final text appears when the clip is done.