TTS Playback Returns [object Object] Instead of Message Audio

[Tom0267]

I'm experiencing an issue with the Text-to-Speech (TTS) feature. When I enable TTS and play a message, the audio plays [object Object] instead of playing the message. Speech-to-Text (STT) works as expected.

Environment Details

• Host OS: Windows 11
• IDE: Visual Studio Code
• Application: LibreChat running in Docker containers

Steps to reproduce

1. Clone LibreChat repository on Windows machine

2. Create/edit .env file using a Windows text editor (VS Code, Cursor, Notepad, etc.)

3. Add OpenAI API key to .env

4. Configure LibreChat with OpenAI TTS in librechat.yaml:

   registration:
     socialLogins: ['github', 'google', 'discord', 'openid', 'facebook', 'apple', 'saml']
     speech:
       tts:
         openai:
           apiKey: '${TTS_API_KEY}'
           model: 'tts-1'
           voices: ['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer']
       stt:
         openai:
           apiKey: '${STT_API_KEY}'
           model: 'whisper-1'

5. Start application with docker compose up

6. Navigate to LibreChat UI and attempt to use Text-to-Speech feature

7. Click the speaker icon on any message

8. Observe error in audio

What browsers are you seeing the problem on?

Chrome