ComfyUI Node
Prompt Rewriter
A ComfyUI node in Prompt Manager with 7 inputs and 1 output.
Prompt Rewriter
- options
- output
◄prompt►
◄seed0►
◄backendCUDA►
◄show_everything_in_consolefalse►
◄keep_mmproj_loadedtrue►
◄stop_server_afterfalse►
CategoryPrompt Manager
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | Enter the prompt you want to embellish | |
| seed | INT | 00–18446744073709550000 | Seed for reproducible generation. |
| backend | COMBO | CUDA | Backend: CUDA (local llama_binaries) or Vulkan (system PATH llama-server) |
| options | OPTIONS | Connect options node to control model and parameters | |
| show_everything_in_consoleopt | BOOLEAN | false | Print system prompt, user prompt, thinking process, and raw model response to the console. |
| keep_mmproj_loadedopt | BOOLEAN | true | Keep mmproj loaded between runs to avoid server restarts (uses more VRAM) |
| stop_server_afteropt | BOOLEAN | false | Stop the llama.cpp server after each prompt (for resource saving, but slower). |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| output | STRING | — |