Audio Transcription

Your API key and audio file are relayed through our server to reach OpenAI's API, but are never stored or logged. The proxy is open source so you can verify this.

OpenAI API Key

Remember key in this browser

Model

gpt-4o-transcribe — Best accuracy and lowest word error rate. ~$0.006/min.

gpt-4o-mini-transcribe — Nearly as accurate at half the cost. ~$0.003/min. Good default for most use cases.

whisper-1 — Original model. Reliable and well-tested, but less accurate than the newer models. ~$0.006/min.

Pricing as of March 2026 — check OpenAI for current rates.

Drop an audio file here or click to upload

MP3, MP4, WAV, FLAC, OGG, Opus*, WebM, M4A — max 25 MB

*Opus files are sent as OGG (same container format)