🎙️

Voice Recorder & Speech to Text

Record your voice and get an instant English transcript. Download audio as WAV or transcript as .txt.

Click to start recording from your microphone

What to expect from browser speech recognition

Speech to Text uses the recognition capability provided by the browser. Accuracy depends on the browser, microphone quality, language selection, background noise, pronunciation, and the service available to that browser.

Because the recognition engine is browser-provided, MangoGlee does not receive or store the recording as a project file. Browser vendors may process recognition audio according to their own terms and privacy policies.

Use a quiet room and speak close enough to the microphone.
Choose the recognition language before starting a recording.
Review names, numbers, and punctuation in the transcript before exporting it.

Features

🎙️Real-time Transcription — As you speak in English, the transcript appears instantly on screen — no waiting, no post-processing.
💾Download Everything — Save your recording as a WAV audio file or export the transcript as a .txt file.
📊Live Waveform — See your voice as a live oscilloscope waveform while recording.
🔒Private Recording — Recording stays on your device. Only the speech recognition stream is processed by Chrome/Edge's built-in API.

How to Use

Click Start Recording and allow microphone access.
Speak in English — your words appear as a live transcript in real time.
Click Stop when done.
Copy the transcript, download it as .txt, or save the audio as WAV.

Frequently Asked Questions

Which browsers support speech-to-text?: Speech recognition requires Chrome or Edge. Firefox does not support the Web Speech API. The recording and WAV download work in all modern browsers.
Is my audio sent to a server?: The recording itself stays on your device. The speech-to-text feature uses Chrome/Edge's built-in Web Speech API, which sends audio data to Google's servers for recognition. MangoGlee never receives your audio.
How accurate is the transcription?: Accuracy is typically very high for English in quiet environments with clear speech. Background noise and accents may reduce accuracy.
Is there a recording time limit?: No hard limit — but very long recordings may use significant memory. For best results, keep individual recordings under 30 minutes.