Skip to main content

API Error Handling

When requests fail, GoingMerry returns standard HTTP status codes accompanied by a structured JSON error body.

1. Error Payload Schema​

All error payloads follow a unified format:

{
  "error": "Error message description detailing what failed"
}

2. Common Status Codes​

400 Bad Request​

The request body contains invalid parameter types or missing fields.

  • Example: Sending a number instead of a string array for chat messages.
{ "error": "invalid messages parameter format" }

404 Not Found​

The requested model configuration tag does not exist locally.

  • Example: Invoking merry run non-existent-model.
{ "error": "model 'non-existent-model' not found, pull it first" }

500 Internal Server Error​

The GPU failed to allocate memory or compile branchless kernels.

{ "error": "out of memory: failed to allocate 4.7GB tensor cache on GPU" }