Audio Upload and On-Device Analysis Are Here

The latest Genderfluent release is here, with new ways to analyze recordings and more control over how voice feedback works. You can now upload prerecorded audio, keep supported analysis entirely on your device, and choose when different types of feedback appear. We’ve also increased the daily real-time formant allowance for free accounts from one minute to five.

Upload prerecorded audio for analysis

You are no longer limited to audio recorded inside Genderfluent. The new Upload Audio feature lets you bring in a prerecorded file and analyze it with the same kinds of feedback you use during regular practice. Click on the upload icon in the lower right corner to try it out.

This can be useful when you want to review a voice memo, revisit an older practice clip, or analyze a recording captured in another app. On the upload screen, pitch is included automatically, and you can choose whether to add formant, volume, and gender analysis before starting.

Because audio upload is a beta feature, we’ll continue refining the experience as more people use it with different file types, devices, and recording conditions. Clear spoken audio with minimal background noise will generally provide the most useful results.

Fully on-device audio analysis

For those who are privacy sensitive or don’t have reliable internet, you can now choose to have all audio analysis performed on-device. Enable this by going to General Settings, then Feedback and privacy, and selecting Local only under Analysis location. However, not all devices will support it since some feedback types like formants are computationally intensive. In those cases, the analysis will either have to occur at the end of the recording or in the cloud. The app will let you know if your device can support on-device analysis for each feedback type.

This gives you a clearer privacy choice without removing flexibility. The Feedback and privacy settings include three analysis-location options:

• Local only keeps analysis on your device.
• Local first tries your device first and can fall back to cloud analysis when necessary.
• Prefer cloud uses cloud processing for supported analysis types, while pitch and volume continue to run locally.

Our Feedback and Privacy Settings guide explains each option, per-analysis overrides, and the device compatibility check in more detail.

Choose when each type of feedback appears

Different types of voice analysis require different amounts of processing power. A device may handle pitch smoothly in real time, for example, while formant or gender analysis works better after the recording has finished.

This release gives you more control over that balance. You can choose Live when available or After recording as your default, then customize the timing for individual analysis types. That means you can keep lightweight feedback live while moving a more demanding model to after-recording analysis.

These controls can help reduce stuttering or interruptions on lower-powered devices. They can also make practice feel less visually busy when you would rather speak first and review the results afterward.

Analysis models now come with the app

The audio analysis model files are now included in the app bundle, which is why the app bundle size increased. Previously, a model would need to be downloaded separately before a feature could run, which could cause problems on slow or unreliable connections.

Five free minutes of real-time formant feedback

We’ve also updated the features available without a subscription. Free accounts now include five minutes per day of real-time formant estimation, up from one minute per day.

That matches the existing free allowance for real-time gender estimation, which is also five minutes per day. The additional time gives you more room to explore how resonance-related measurements change across words, vowels, and practice sessions. If formants are new to you, our introduction to vocal formants explains what F1, F2, and F3 measure and why they can be useful alongside pitch.

Bug fixes and performance improvements

This release also includes smaller fixes and performance improvements across the app. Together with bundled model files and more flexible analysis settings, these changes should make recording and reviewing your voice feel smoother and more reliable.

---

As always, thank you for your support and feedback! We’re excited to continue improving Genderfluent and helping you on your voice journey. If you have any questions or suggestions, please reach out to us at info@genderfluentapp.com.