OpenAI’s release notes for October 6, 2026 say ChatGPT can now take an audio file and turn it into a transcript, a summary, or answers about what was said. For anyone who sits through weekly meetings or records lectures, that turns a recording from something you replay into something you can question. The catch is the price of entry, because ChatGPT audio uploads are limited to paid plans and workspaces, and the Free plan is left out.
TL;DR: ChatGPT audio uploads let paid users and workspaces turn a meeting, interview, or lecture recording into a transcript, summary, or follow-up email draft. Files can be up to 512 MB in formats including MP3, WAV, and M4A. The Free plan does not get it. OpenAI warns that transcripts can contain errors and speaker labels may be unreliable, so check anything important against the original recording.
What ChatGPT audio uploads can do
You attach a supported audio file to a conversation and describe what you need, according to the release note. That can be a transcript, a summary of the recording, or questions about its contents.
OpenAI’s own examples are practical ones, such as turning a meeting, interview, or lecture into structured notes or a follow-up email draft. The help center adds two prompts: ask for a transcript and then follow-up questions, or summarize a meeting recording with decisions, open questions, and next steps.
Picture a forty minute call where three people walked away with three different ideas of what was agreed. Instead of scrubbing back through the audio, you ask which decisions were made and who owns each one.
The details below come from OpenAI’s help center article on uploading files and audio.
| Detail | What OpenAI says |
|---|---|
| Plans | Paid ChatGPT subscriptions and workspaces, including Enterprise |
| Free plan | Not available at this time |
| File size | Up to 512 MB |
| Upload cap | 80 files every 3 hours |
| Platforms | Web and supported mobile apps |
| Long recordings | May be processed in smaller sections, and very long ones may time out |
The useful question is what your file has to look like to get through.
Formats, size limits and who gets access
The help center lists WAV, MP3/MPEG, OGG/OGA, audio-only WebM, PCM, FLAC, AAC, M4A, and audio-only MP4. WebM and MP4 files identified as video are not supported, so a video file falls outside the list even if it carries spoken audio.
Audio files can be up to 512 MB. Longer recordings may be processed in smaller sections when Data Analysis is available, and OpenAI describes processing as best-effort.
Access is the sharper edge. Audio uploads come with paid subscriptions and workspaces, and OpenAI adds that availability may vary by workspace settings, region, client version, and model.
The same page sets a general cap of 80 file uploads every 3 hours, and Free users are limited to 3 file uploads per day. Audio is off the table for them either way. That leaves the question of how far to trust what comes back.
Where transcripts can go wrong
OpenAI states that transcripts may contain errors, that speaker identification may be unreliable, and that performance may vary across languages.
That is the detail that decides whether this saves time or creates a new kind of mistake. If a summary credits the wrong person with a decision, the error rides along into the follow-up email you send.
The help center’s own advice is to review important details against the original recording. In practice, the recording stays in your workflow even after the notes exist.
Privacy is the other thing worth settling before you upload someone else’s voice.
What to check before uploading a recording
File retention depends on storage location and workspace policies, per the help center. For consumer services, data usage follows OpenAI’s standard policies, while content submitted to business offerings such as the API and ChatGPT Enterprise is not used to improve model performance.
The page does not say how uploaded audio is deleted, or whether Temporary Chat changes anything for audio, so treat both as unanswered for now.
ChatGPT has been taking in more personal material lately, from the ChatGPT Health tab to linked finances. A recording of a client call or a family conversation is a different kind of input from a PDF, and our guide to phone privacy settings covers the wider cleanup.
