Skip to main content

​
Overview

The health check endpoint allows you to verify that all required Unmute backend services are operational. This endpoint checks the status of the TTS (Text-to-Speech), STT (Speech-to-Text), LLM (Language Model), and Voice Cloning services.

​
Endpoint

curl https://api.unmute.example/v1/health

​
Response

The endpoint returns a JSON object with the status of each service:
​
tts_up
boolean
required
Indicates whether the Text-to-Speech service is operational
​
stt_up
boolean
required
Indicates whether the Speech-to-Text service is operational
​
llm_up
boolean
required
Indicates whether the Language Model service is operational
​
voice_cloning_up
boolean
required
Indicates whether the Voice Cloning service is operational
​
ok
boolean
required
Overall health status. Returns true if TTS, STT, and LLM services are all operational. Note that voice cloning is not required for the server to be considered healthy.

​
Response Example

{
  "tts_up": true,
  "stt_up": true,
  "llm_up": true,
  "voice_cloning_up": true,
  "ok": true
}

​
Error Response

If any critical service is down, the response will indicate which services are unavailable: 
{
  "tts_up": true,
  "stt_up": false,
  "llm_up": true,
  "voice_cloning_up": false,
  "ok": false
}

​
Status Codes

​
200
status code
Successfully retrieved health status (returned regardless of whether services are healthy or not)

​
Caching

This endpoint implements a 0.5-second TTL cache to prevent excessive health checks from overwhelming the backend services.

​
Usage Notes

  • The ok field only considers TTS, STT, and LLM services as critical
  • Voice cloning service status is informational but does not affect the overall ok status
  • Health checks are performed asynchronously in parallel for optimal performance
  • This endpoint should be called before initiating a WebSocket connection to /v1/realtime

Build docs developers (and LLMs) love

Get started for freeTalk to us