Audio to Text Converter — Free Online Transcription

Drag any audio file — MP3, WAV, M4A, FLAC, OGG — and get back searchable text with timestamps. 99.5% AI accuracy on sermons, interviews, podcasts, lectures, and meetings. First 10 minutes are free; everything after is $0.006/min.

What this is

Audio-to-text conversion (also called speech-to-text or transcription) uses a neural network to turn recorded speech into written words. Modern AI models like OpenAI Whisper achieve near-human accuracy on clear voice recordings at a tiny fraction of the cost of professional human transcription.

This page is the universal entry point — upload anything from a podcast MP3 to a lecture WAV to an iPhone Voice Memo M4A, and the same backend handles it. If you have a specific format in mind we have dedicated guides for MP3, WAV, M4A, MP4 video, and YouTube links.

Step-by-step: audio file to text

1. 1

   Open the upload tool

   Go to /transcribe in the nav. No signup needed for your first 10 minutes of audio.

2. 2

   Drop the audio file in

   Drag-drop from Finder, Explorer, Dropbox, or Google Drive. Up to 25MB on the free tier; 500MB per file on Pro. Accepted: MP3, WAV, M4A, AAC, FLAC, OGG, OPUS, AIFF.

3. 3

   Pick Standard or Premium

   Standard ($0.006/min, OpenAI Whisper) is the right call for most single-speaker recordings. Premium ($0.02/min, ElevenLabs Scribe) adds speaker diarization — labels who said what. Pick Premium if you're transcribing a panel, interview, or roundtable.

4. 4

   Let the engine run

   Processing takes ~10% of audio length. A 45-minute file transcribes in about 5 minutes. You can close the tab; results land in your dashboard and an email goes out when done.

5. 5

   Review the transcript in the browser

   Hit play in the inline player — each word is clickable and seeks the audio to that timestamp. Edit typos directly. Search-find works for jumping to specific phrases.

6. 6

   Download the output

   Choose .txt (plain text), .srt (subtitle format), .vtt (HTML5 captions), or .docx (Word document with formatted timestamps). Pro accounts also get JSON with word-level timestamps for custom processing.

Supported audio formats

Tips that lift accuracy 1–3%

• Mic placement is everything. A $20 lavalier within 6 inches of the speaker beats a $2,000 condenser at 6 feet, every time. The biggest accuracy gains come from before the recording, not after.
• Trim silence at the start and end with Audacity (Effect → Truncate Silence) before upload. Smaller files upload faster and you don't pay to transcribe dead air.
• Normalize the audio if levels are quiet. Audacity Effect → Normalize → -1.0 dB. This makes a measurable accuracy difference on under-gained recordings.
• Mono, not stereo, for voice. The model collapses to mono internally, but a properly summed mono export skips one step and gives marginally cleaner input.
• Two voices speaking over each other? Use Premium tier. Standard treats overlapping speech as one stream and confuses the lattice.

The audio-to-text workflow

Audio-to-text pricing vs alternatives

Pricing as of early 2026. Rev AI $0.25/min; Rev human $1.50/min. Otter Pro $16.99/mo ÷ 1,200 included min × 45 min. HappyScribe AI ~$0.20/min.

Audio to text FAQ

Related