Prompt an AI
POST/prompt-svc/prompt
Sends a prompt and waits for a response if sync is true. If sync is false, adds the prompt to the queue and returns immediately.
Prompts can be used for text-to-text
, text-to-image
, image-to-image
, and other types of generation.
If no model ID is specified, the default model will be used (see Model Svc
for details). The default model may or may not support the requested generation type.
Prompting Modes
- High-Level Parameters: Uses predefined parameters relevant to
text-to-image
,image-to-image
, etc. This mode abstracts away the underlying engine (e.g., LLaMA, Stable Diffusion) and focuses on functionality. - Engine-Specific Parameters: Uses
engineParameters
to directly specify an AI engine, exposing all available parameters for fine-tuned control.
Permissions Required: prompt-svc:prompt:create
Request
Responses
- 200
- 400
- 401
- 500
OK
Invalid JSON
Unauthorized
Internal Server Error