Excellent features are coming to the Gemini app! It's now possible to upload audio files, even for free

The Gemini app finally received support for uploading audio files, a feature users have been requesting from the very beginning. The interesting part is that it has been available in the API for two years and worked in AI Studio since the release of model 2.5 Pro. It just… wasn’t in the app itself.

According to Google, it was the most requested feature ever. Which makes sense – who wouldn’t want AI to transcribe an hour-long meeting recording or analyze a podcast?

How to do it? Just click the plus sign

Using it is as simple as can be. On mobile, open the menu with the plus sign and select “Files”, then on the web, “Upload files”. The app will accept virtually any audio format – MP3, M4A, WAV, and many others. Then just type what you want to do with the recording.

But here comes the first catch. Free users can upload a maximum of 10 minutes of audio. If you have a Google AI Pro or Ultra subscription, the limit increases to 3 hours. That’s a difference worth considering, especially if you regularly work with longer recordings.

What all can Gemini handle?

Audio files aren’t the only thing you can upload to Gemini. The app previously supported videos – 5 minutes for free or an hour with a subscription, maximum 2 GB. More interesting is the support for developers:

• Code folder or GitHub repository – up to 5000 files, up to 100 MB
• ZIP files – maximum 10 files inside
• Other formats – individual files up to 100 MB

Overall, you can upload up to 10 files of any format to a single chat. That’s not bad, even if competitors sometimes offer more.

Why the hell did it take so long?

Users on Reddit are rightly asking – why was a feature available in the API for two years not in the app? The answer is probably prosaic. Google needed to ensure sufficient computational capacity for millions of users. It’s one thing to offer a feature to developers via an API, another to scale it for the masses.

Moreover, as one user noted, the problem wasn’t with the model itself, but with the app’s user interface. Google simply needed time for implementation. Which is a bit shameful for a company that boasts about its AI capabilities at every turn.

First impressions? Surprisingly good

According to initial feedback, the feature works surprisingly accurately. One user tested the analysis of a song with a lot of distortion, and Gemini was able to recognize instruments and lyrics, albeit with occasional humorous mishearings. For transcribing regular recordings or analyzing podcasts, it should be more than sufficient.

Of course, it doesn’t work for everyone right away – some users report issues with file processing. However, this is unfortunately standard for Google’s gradual rollout of new features. The main thing is that the feature finally exists. Even if it’s two years late.

At the time of writing, it wasn’t available on the editorial Pixel 10 Pro, but that will reportedly change in a few hours.

Will you use the option to upload audio files to Gemini?

Source: 9to5google