Audio to text,
with your privacy intact.
Convert audio to text from a short voice note or spoken recording. Create a draft on your device, correct it, and copy or download a TXT file. Free to use, with no account or audio uploads.
Choose audio to transcribe
A short voice note or spoken recording, saved on your device.
Up to 5 minutes · 25 MB per file
Creates an editable text draft and approximate subtitle cues. No speaker labels, translation, or live recording.
The first run downloads about 70–80 MB of speech recognition resources. Cached resources may reduce later downloads. A computer is recommended; accuracy and cue timing vary with speech and recording quality.
How to convert audio to text
Use this audio to text converter for personal notes, short spoken ideas, or a brief interview excerpt. Clear speech from one person is a better starting point than overlapping voices. Review names, numbers, and anything important before using or sharing the draft.
- Recording
- Up to five minutes and 25 MB. Common WAV, MP3, and M4A recordings can be selected; decoding depends on your browser. If a file cannot open, convert it to WAV first.
- Languages
- English, Spanish, French, German, Italian, Portuguese, Chinese, Japanese, Korean, Hindi, Arabic. Choose the language spoken in the recording. Accuracy varies by language, accent, noise, and recording quality.
- Text output
- An editable plain-text draft, plus approximate editable cues for SRT and VTT subtitle downloads. Speaker labels, translation, and meeting summaries are not provided.
Choose a short recording and select its spoken language. Try the English sample to explore the tool.
Choose Transcribe audio. The first use downloads about 70–80 MB of recognition resources. Keep the tab open while your device converts audio to text.
Review the transcript and cue timing. Edit and download TXT, SRT, or VTT, and save your work before refreshing or leaving.
Hear the recording and read a sample transcript
This synthetic English recording was transcribed with this tool. The spoken sentence uses “nine”; the draft below writes it as “9”. Review the wording and numbers in your own transcript too.
The meeting starts at 9. Please bring your notes.
Download the original recording · Download the sample TXT. This is one short example, not an accuracy guarantee. Try the spoken sample above to generate a draft on your device.
Audio to text questions
Is this audio to text converter free?
Yes. There is no account or paid quota. Each recording is limited to five minutes and 25 MB, and processing uses your own device. Copy the reviewed text or download TXT, SRT, or VTT before leaving the page.
How accurate is audio to text transcription?
Accuracy depends on the language and recording. Noise, overlapping voices, accents, names, and technical terms can cause mistakes or missing words. The small local model is best used for drafts; check important information against the original.
Does speech to text mean live dictation or translation here?
Here, speech to text means transcribing a saved recording in its selected spoken language. There is no microphone recording, live dictation, or translation into another language.
Can I identify speakers or create subtitles?
Speaker labels are not available. The tool can return approximate speech cues that you can edit and download as SRT or VTT. Review their wording and timing against the recording before publishing. Some speech may have no usable cue timing, so TXT can be the better option.
Can I transcribe a long lecture or a whole meeting?
Each file must be no more than five minutes and 25 MB. Split a longer recording in an audio editor and process short excerpts, or choose a service designed for long recordings. Speaker labels and meeting summaries are not provided.
Does audio to text work offline or on a phone?
The first run needs internet to download recognition resources. Your browser may cache them, but storage limits or clearing browsing data can remove them. Your audio is processed locally. A computer with a recent Chrome or Edge browser is recommended; phones may run out of memory.
What is downloaded, and who receives my audio?
This site supplies recognition resources and Hugging Face supplies the Whisper Tiny model. They receive ordinary resource requests, not your selected recording. Audio and transcripts stay in this tab until you save or copy them.