Select CUDA GPU device

PUT

/api/v1/system/gpu-device

Switch the active CUDA device. Unloads both TTS and ASR models for device migration.

Query Parameters

index*Index

CUDA device index

Successful Response

application/json

success*Success

Whether the operation succeeded

changed*Changed

Whether the device was actually changed

needs_reload?Needs Reload

Whether the TTS model needs to be reloaded on the new device

Defaultfalse

message?|

Optional status message

Example Request

curl -X PUT 'http://127.0.0.1:8100/api/v1/system/gpu-device?index=0'

Example Request

curl -X PUT 'http://127.0.0.1:8100/api/v1/system/gpu-device?index=0'

Example ResponseSchema

{
  "success": true,
  "changed": true,
  "needs_reload": false,
  "message": "string"
}

{  "success": true,  "changed": true,  "needs_reload": false,  "message": "string"}

List CUDA GPU devices GET

Return available CUDA devices with names/VRAM and the currently selected index.

Check if setup is required GET

Check whether first-run setup is needed (missing backend, Python upgrade, etc.).