ComfyUI Node
LM Studio Model
A ComfyUI node in LM Studio with 8 inputs and 1 output.
LM Studio Model
- model
◄model<no models found>►
◄manual_model_id►
◄unload_after_usetrue►
◄temperature0.15►
◄max_tokens-1►
◄seed0►
◄image_max_size1024►
◄use_rest_apifalse►
CategoryLM Studio
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| model | COMBO | <no models found> | Model key discovered from the LM Studio SDK. Pick a listed model or use manual_model_id. |
| manual_model_id | STRING | Optional manual model identifier if your LM Studio instance is not returning it via /models or runs on a different base URL. | |
| unload_after_use | BOOLEAN | true | Unload the model after queries to free up memory. Uses LM Studio SDK for proper model management. |
| temperature | FLOAT | 0.150–2 | Sampling temperature. Higher is more random; lower is more deterministic. |
| max_tokens | INT | -1-1–32768 | Maximum new tokens to generate for the assistant reply. Use -1 for unlimited (recommended for thinking models). |
| seed | INT | 00–2147483647 | Optional seed for deterministic sampling if supported by LM Studio. Use 0 to disable. |
| image_max_size | COMBO | 1024 | Maximum edge size for input images (keeps aspect ratio). Images larger than this are downscaled before encoding and sending to LM Studio. |
| use_rest_api | BOOLEAN | false | Use the OpenAI-compatible REST API instead of the LM Studio SDK for predictions. Required for proper thinking model support (separates reasoning from content). |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| model | LMSTUDIO_MODEL | — |