ComfyUI Node
NInfer - Chat
Generates text, optionally conditioned on reference images.
NInfer - Chat
- client
- options
- history
- image_1
- image_2
- image_3
- image_4
- Response
- Reasoning
- History
◄prompt►
◄model►
◄force_regentrue►
◄languagezh►
CategoryNInfer
Inputs (11)
| Name | Type | Default | Description |
|---|---|---|---|
| client | NINFER_CLIENT | From a NInfer client node | |
| prompt | STRING | The user message | |
| model | STRING | Leave empty to use the first model the server reports | |
| force_regen | BOOLEAN | true | Always call the model, even when nothing changed |
| language | COMBO | zh | Language for the preview and status text this node prints. ComfyUI's own node labels follow the interface language. |
| optionsopt | NINFER_OPTIONS | Settings chain from the option nodes | |
| historyopt | NINFER_HISTORY | Conversation so far | |
| image_1opt | IMAGE | Reference image. Connect them in order; each slot adds about 256 prompt tokens at 512px | |
| image_2opt | IMAGE | Reference image. Connect them in order; each slot adds about 256 prompt tokens at 512px | |
| image_3opt | IMAGE | Reference image. Connect them in order; each slot adds about 256 prompt tokens at 512px | |
| image_4opt | IMAGE | Reference image. Connect them in order; each slot adds about 256 prompt tokens at 512px |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| Response | STRING | Visible reply text |
| Reasoning | STRING | Hidden thinking, when the model produced any |
| History | NINFER_HISTORY | Conversation including this turn |