Skip to content

[BUG] SpeechToText does not return consistent text during recognition (i.e. RecognitionResultUpdated) #2982

@AdrianPell

Description

@AdrianPell

Is there an existing issue for this?

  • I have searched the existing issues

Did you read the "Reporting a bug" section on Contributing file?

Current Behavior

Using OfflineSpeechToText with partial results requested produces different results on Android and iOS from the RecognitionResultUpdated event:

  • On Android, each event includes the entire text which has been recognized
  • On iOS, each event only includes new text recognized since the previous event

The difference in this behavior is unexpected since the basic expectation from the toolkit is that behavior is the same on all platforms.

Android result:

Image

iOS result:

Image

Expected Behavior

Consistent behavior - preferably the current iOS behavior - to only return the latest text (to be added to earlier recognized text.

Steps To Reproduce

  1. Open the repo in Visual Studio 2026 with the MAUI workload installed, and connected to appropriate test devices for Android and iOS
  2. Build and deploy to an Android device
  3. Run by pressing the "Start Listening" button and speak into the microphone
  4. The display shows the various stages of recognition, with the whole text to date being displayed
  5. Repeat for iOS, showing only the incremental recognition text

Link to public reproduction project repository

https://github.com/AdrianPell/MauiSpeechToTextBug/tree/5349de193e2d3d8c68d4394706baefc4b804ec73

Environment

- .NET MAUI CommunityToolkit: 13.0
- OS: iOS 18.6.2 or Android 14.0
- .NET MAUI: 10.0.11

The same behavior has been observed in .NET 9

Anything else?

This is the same as #1723 which was closed as "needs reproduction"

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions