A PRACTICAL AUDIO GUIDE
ChatGPT Audio Editor: Change a Recording with a Prompt
You know how you want the recording to sound. You should be able to say it. GPTAudioEditor turns everyday instructions into changes to your existing audio, then gives you a version you can play and download.
Upload audio and start editingStart with the recording you already have
Upload a voice note, podcast excerpt, interview or music clip. MP3, WAV, M4A, AAC, FLAC, OGG, Opus, AIFF, WMA and WebM are accepted inputs. The limit is 100 MB and 20 minutes, and your browser must have enough memory to decode the file. Start with a short section when you are experimenting.
Describe the outcome in your own words
You do not need to ask for a highpass filter or a compressor. Say “make my voice clearer,” “remove the first five seconds,” or “fade out the ending.” Give exact times when a cut needs to happen at a particular moment. For a complicated recording, make one meaningful change, listen, then add a follow-up.
Hear the difference before you keep it
The edited recording becomes the current version. Switch between Original and Edited to compare, and use Versions to go back. An AI plan is not proof of a better result: listen for changes to words, breaths, music transients and the ending. Download only after the result sounds right to you.
What “ChatGPT audio editor” means here
GPTAudioEditor is an independent audio editing product, not ChatGPT or an OpenAI service. Its editing assistant uses selected models through OpenRouter, and the browser performs the edits. It does not generate replacement speech, clone voices or separate instruments into stems. The assistant explains when a request needs a different tool.
A prompt to try
Make this voice recording clearer and a little more even in volume. Keep the voice natural and do not cut any words.
Check that quiet words remain audible and that the voice has not become harsh. Compare at a similar listening volume.
Common questions
Do I need a ChatGPT subscription?
No. GPTAudioEditor has its own account and audio credits. A ChatGPT subscription does not provide credits here.
Does the AI hear the whole file?
No. By default it receives your request, timing and detected silence ranges. If you enable listening, it receives at most a 30-second preview beginning at your selected time, or the start of the file.