VAPI TTS is down, if I choose anything else it wor...
# support
f
VAPI TTS is down the agent will not speak back, if I change it to openai or any other TTS the agent start speaking, the only thing that works is the first message, and the voice ID is spenser.
v
Suggestion: Double-check that your Assistant ID and Phone Number ID are correctly set in your VAPI tool configuration, as missing or incorrect values can cause the agent not to speak even if the TTS provider is selected. Example configuration:
Copy code
json
{
  "assistantId": "your-assistant-id",
  "phoneNumberId": "your-phone-number-id"
}
If these are set and the issue persists, try switching to a different voice ID to see if the problem is specific to "spenser" ([source](https://discord.com/channels/1211482211119796234/1370485222214336564)). Source: - [Discord VAPI TTS troubleshooting thread](https://discord.com/channels/1211482211119796234/1370485222214336564)
f
call id 019a7ffb-b2ee-7ee4-aa16-565316a4e2d5
seems like an issue with Spenser voice only
e
Best practice is to configure a fallback TTS and STT model. This can only be done via API call, cannot be done in the dashboard.
c
Hi Freedom, To integrate a custom text-to-speech (TTS) system with VAPI, you'll need to set up webhook authentication, build a compatible TTS server, and ensure your system can handle VAPI's audio requirements. 1. **Authentication Setup**: Use custom credentials for secure communication by creating them in your VAPI dashboard ([see details here](https://docs.vapi.ai/customization/custom-voices/custom-tts)). 2. **Build TTS Server**: Your server should: - Handle POST requests from VAPI with text and audio specs. - Generate audio in raw PCM format and return it in the response. 3. **Response Structure**: Ensure your server responds with: - HTTP 200 status -
Content-Type: application/octet-stream
- Raw PCM audio data 4. **Audio Specifications**: Follow these guidelines for your TTS system: - Raw PCM format (no headers) - Mono channel - 16-bit signed integer - Little-endian byte order - Matching sample rate with VAPI's request (e.g., 8000, 16000, or 24000 Hz) 5. **Testing**: Create test calls using VAPI’s API and monitor requests to ensure your setup is correct. For a complete guide on setting up custom TTS with VAPI, check the [documentation here](https://docs.vapi.ai/customization/custom-voices/custom-tts). If you face any connectivity issues, verify the network communication between VAPI and your TTS server, and consider increasing the
timeoutSeconds
if processing takes longer.