Transcribe audio to text for free
Transcribe audio to text for free with local AI. Turn recordings into editable text and export TXT, SRT or VTT in your browser, without uploading audio.
Your audio stays on your device. Speech recognition runs in your browser.
Drop an audio file here
MP3 · WAV · M4A · AAC · OGG · FLAC — decoding support depends on your browser
A smaller model for a first try. Model weights download roughly 40–150 MB, depending on processing mode. It may miss words, especially in noisy or multilingual recordings.
Auto will use CPU because the browser cannot confirm a suitable discrete GPU. You can still choose GPU manually.
Auto mode detects the language from the first section with speech and uses it for the rest of the recording. If the result is uncertain, choose the spoken language manually. The model and runtime download on first use. Starting again replaces the current transcript.
Chinese recognition and writing style are separate. Choose Simplified or Traditional Chinese for display, copying and exports. You can switch after transcription without running recognition again.
How to transcribe audio online
Choose your recording
Select an audio file from your device, or continue after downloading audio with FlowPick. Your recording is read locally.
Recognize the speech
Choose the spoken language and a model, then start. The model downloads on first use. Longer recordings are processed in sections.
Review and export
Listen at each timestamp and correct the transcript. Save plain text for notes, or SRT and VTT files for subtitle workflows.
Audio to text without the upload
Convert MP3, WAV and M4A audio into text
This audio to text converter runs Whisper speech recognition inside your browser, so recordings are not sent to a transcription API. Choose the spoken language or let auto mode detect it, pick the lightweight or the standard model, then correct the text against the audio and export it. First use downloads the model, and recognition speed depends on your device — GPU mode is optional when the browser exposes it.
MP3 to text · WAV to text · voice memo to text · podcast transcript · interview transcription · SRT and VTT subtitles
Which audio files can be transcribed?
Any file your browser can decode can be used as input. Recognition quality depends on the speech itself, not on the container, so converting a file does not automatically improve the transcript.
| Audio input | Typical source | What to know |
|---|---|---|
| MP3 | Podcasts, exports from phones and other apps | Small files with broad decoding support |
| WAV | Recorders and edited or converted audio | Large, but the most predictable input for decoding |
| M4A / AAC | iPhone voice memos and mobile recordings | Common on mobile; decoding depends on the browser |
| OGG / Opus | Web and voice-chat exports | Efficient for speech; browser support varies |
| FLAC | Lossless archive recordings | Larger files need more memory while decoding |
From recordings to readable text
Interview transcripts
Find passages in a recorded conversation and correct names or technical terms while listening. Speaker labels are not generated in this version.
Podcast notes
Turn a podcast audio file into text you can search, quote and edit. The output is a transcript rather than an automatic summary.
Lecture and voice recordings
Keep a text copy of spoken explanations. Clear speech and a matching language setting help recognition; verify important details yourself.
Subtitle preparation
Export timestamped SRT or WebVTT subtitles. Edit the wording and check the timing in your video editor before publishing.
Supported audio and practical limits
MP3, WAV, M4A, AAC, OGG and FLAC support depends on the codec and browser decoder. Recognition runs on your device, so long recordings can take considerable time and battery power. The tool processes overlapping sections to limit decoded-audio memory, but the model still needs memory of its own. Small models can mishear accents, quiet speech, music and overlapping speakers. Review the text and timing before use. Media and transcripts are not uploaded by this tool; model downloads require a network connection.
Audio transcription questions
Can I convert MP3 to text?
Yes, if your browser can decode the MP3 file. Choose the recording, select its spoken language and transcribe it. You can edit the result and export TXT, SRT or VTT.
Is my recording uploaded?
No. File reading and speech recognition take place in your browser. The tool downloads model files and a runtime, but does not send your audio or transcript to a transcription service.
Can I use this offline?
First use requires network access to load the tool, runtime and model. Cached model files may be reused, but this version does not guarantee fully offline operation. Browsers can remove cached files.
Which languages can I transcribe?
Auto mode detects the language from the first section with speech and keeps it for the recording. Choose a language manually for mixed-language or uncertain recordings. Chinese output can be simplified or traditional; this changes the writing style and does not translate the speech.
Does it identify different speakers?
This version produces timestamped text without speaker identification. You can add speaker names while editing the text.
What happens if I cancel?
The background task stops and releases its resources. Completed sections remain available for editing and export. Starting another run begins from the start and replaces the current text.
Is this audio to text tool free?
Yes. Transcription is free and does not require an account. Recognition runs on your own device, so the only network use is the model and runtime download on first run.
Can I turn a voice memo into text?
Yes, as long as your browser can decode the memo. M4A and AAC voice memos from phones usually work. Select the file, set the spoken language, then edit the transcript and export it as TXT, SRT or VTT.
Can I transcribe the audio of a video file?
Use Video to Text to transcribe the primary audio track directly, without extracting it first. If you also need a separate audio file, use Audio Extractor and continue with that file.
How accurate is the transcript?
The lightweight and standard models are useful, not exact. Clear speech, one main speaker and a correctly chosen language give the best results. Accents, background noise, music, crosstalk and technical vocabulary may be misheard, so review the text before relying on it.
How long can the recording be?
There is no fixed upload limit, because the file stays on your device. Long recordings are processed in overlapping sections and can take considerable time, memory and battery power. For very long files, trim the audio to the part you need first.
Can I create subtitles for a video?
Yes. Correct the sections, export SRT or WebVTT, then import the file into your video editor or upload platform. Timestamps are approximate, so check the timing and wording on the timeline before publishing.
More audio tools
AI Vocal Separator
Separate vocals and instrumentals for free with local AI. Preview both tracks and download WAV files in your browser, without uploading your audio.
Audio Noise Remover
Reduce background noise in voice recordings for free. Preview, compare and download a WAV file with local AI in your browser, without uploading audio.
Video to Text
Convert video speech to text for free in your browser. Edit transcripts and export TXT, SRT or VTT subtitles with local AI, without uploading your video.
Audio Converter
Change a recording to WAV or MP3 before transcribing it.
Audio Extractor
Pull the audio track out of a video, then send it straight to text.
Audio Trimmer
Cut a long recording down to the part you want to transcribe.
Audio Volume Booster
Make quiet speech louder first to give recognition a better chance.
Audio Downloader
Save a direct audio resource from a web page, then transcribe it locally.