issue: Call Mode TTS audio is not played on mobile browsers but works on desktop

Check Existing Issues

• I have searched for any existing and/or related issues.
• I have searched for any existing and/or related discussions.
• I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!).
• I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

v0.10.2

Ollama Version (if applicable)

Operating System

iOS 26.5.2, Android 16.0 (Baklava), Windows 11

Browser (if applicable)

Windows 11: 150.0.7871.125, iOS 26.5.2: 150.0.7871.113, Android 16.0 (Baklava): N/A

Confirmation

• I have read and followed all instructions in README.md.
• I am using the latest version of both Open WebUI and Ollama.
• I have included the browser console logs.
• I have included the Docker container logs.
• I have provided every relevant configuration, setting, and environment variable used in my setup.
• I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
• I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
• Start with the initial platform/version/OS and dependencies used,
• Specify exact install/launch/configure commands,
• List URLs visited, user input (incl. example values/emails/passwords if needed),
• Describe all options and toggles enabled or changed,
• Include any files or environmental changes,
• Identify the expected and actual result at each stage,
• Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

When using Call Mode on a mobile device, Open WebUI should:

1. Record and transcribe the user's speech completely.
2. Generate the assistant response.
3. Automatically convert the response to speech.
4. Play the generated TTS audio through the mobile device.

The behavior should be the same as on desktop Chrome, where the assistant correctly speaks its response aloud.

Actual Behavior

On mobile devices, Call Mode records and transcribes the user's speech and the assistant generates a visible text response.

However, the generated TTS response is not played. The mobile device remains completely silent.

The same Open WebUI deployment, model, TTS configuration, and user account work correctly on desktop Chrome, where the assistant speaks the response aloud.

Additionally, the beginning of the spoken input is sometimes cut off on mobile devices. For example:

Spoken: "Wie wird das Wetter heute?"

Transcribed: "Die geht das hah te?"

The loading indicator remains active indefinitely until Call Mode is stopped manually. This is only in the Mobile Version.

Steps to Reproduce

1. Deploy Open WebUI in a Kubernetes cluster using the official Docker image.

2. Configure and enable Speech-to-Text and Text-to-Speech in the Open WebUI audio settings.

3. Open the Open WebUI instance in Chrome on a desktop PC.

4. Log in and open any chat.

5. Start Call Mode.

6. Speak a short sentence, for example:

   "Wie geht es dir heute Morgen?"

7. Wait for the assistant response.

8. Confirm that on desktop Chrome:

   • The speech is transcribed.
   • The assistant response is generated.
   • The response is played aloud through TTS.

9. Open the same Open WebUI instance on a mobile device.

10. Test with:

    • Chrome on iOS
    • Chrome on Android

11. Allow microphone access when prompted.

12. Log in with the same user account.

13. Open the same chat and select the same model.

14. Start Call Mode.

15. Speak the same sentence:

    "Wie geht es dir heute Morgen?"

16. Wait for the assistant response.

17. Observe that:

    • The speech is transcribed.
    • The assistant response is displayed as text.
    • No TTS audio is played.
    • The mobile device remains silent.

18. Repeat the test several times.

19. In some attempts, also observe that the beginning of the spoken sentence is cut off or changed during transcription.

20. The loading indicator remains active until Call Mode is stopped manually.

Logs & Screenshots

Not needed.

Additional Information

The issue appears to be specific to mobile browser audio playback in Call Mode.

The following parts still work on the affected mobile devices:

• Open WebUI loads correctly.
• Login works.
• Microphone permission is granted.
• Speech input is recorded.
• Speech-to-Text returns a transcription.
• The prompt is sent successfully.
• The assistant response is generated and displayed as text.

Only the automatic TTS playback is missing.

The same Open WebUI instance, user account, model, and TTS configuration work correctly in desktop Chrome.

Affected devices tested:

• iPhone 15 with Chrome on iOS
• Android device / Android emulator on Android Studio with Chrome

Working client:

• Desktop PC with Chrome