ChatGPT can now transcribe audio files that you upload directly into a conversation, and while that might sound like a small update, I can already think of plenty of situations where this could be useful. OpenAI rolled out audio file uploads to ChatGPT on 6 October 2026. You can attach an existing recording and ask ChatGPT to transcribe it, summarise the conversation, extract important information, create meeting notes, or even prepare a follow-up email based on what was discussed.
I’ve had Zoom meetings in the past that were recorded but never transcribed. If I needed to go back and find a particular discussion, decision, or piece of information later, the only practical option was to scrub through the recording and try to find it manually.
With ChatGPT’s new audio upload support, I could instead upload that old recording and ask questions such as “What did we decide about the launch date?”, “Summarise the main action items”, or “Find where we discussed pricing”.
For me, that’s probably one of the most useful scenarios for this feature but it doesn’t necessarily have to be a Zoom recording. OpenAI currently supports WAV, MP3/MPEG, OGG/OGA, audio-only WebM, PCM, FLAC, AAC, M4A and audio-only MP4 files. Audio uploads can be as large as 512MB, although OpenAI warns that very long recordings may be processed in sections and could potentially time out.
So if you have recordings of meetings, interviews, presentations or lectures sitting around on your computer, they can potentially become searchable information rather than files you rarely open again. Or convert songs into lyrics, and movie audio into transcripts.
There are already plenty of tools that can convert speech into text, so transcription by itself isn’t particularly exciting in 2026. The advantage here is what you can do afterwards.
Once ChatGPT understands the recording, you can continue asking questions about it. You could ask it to turn a one-hour meeting into five key points, separate decisions from general discussion, create a list of tasks, draft an email for everyone who attended, or explain a section you didn’t understand.
For journalists and content creators, uploading an interview and then asking ChatGPT to identify particular topics could also save a significant amount of time when going back through a recording.
OpenAI already has a ChatGPT Record feature that can capture and summarise meetings and voice notes, but that currently has its own platform and workspace restrictions. Audio uploads address a slightly different problem because the recording can already exist before you bring it into ChatGPT.
At launch, audio uploads aren’t available to Free users. OpenAI says the feature is available with paid ChatGPT subscriptions and workspaces, with availability also depending on region, workspace settings, app version and the model being used.
You also shouldn’t assume every transcript will be perfect. OpenAI specifically warns that transcription errors can occur and that identifying individual speakers may not always be reliable. Performance can also vary between languages.
For important meetings, interviews or anything where exact wording matters, I would still check the generated transcript against the original recording. Even with those limitations, though, I think this is one of those ChatGPT updates that may turn out to be much more useful in everyday life than it first appears.
Being able to upload a meeting that happened months ago, get a transcript within ChatGPT and then effectively have a conversation with that recording removes a lot of tedious manual work. And if, like me, you have old Zoom recordings that were never transcribed in the first place, they may suddenly be useful again.
