ComfyUI Node
Prompt Studio (TJ)
A ComfyUI node in ✨ TJ_Node/LLM with 30 inputs and 2 outputs.
Prompt Studio (TJ)
- image
- clip
- prompt
- thought_process
◄get_name_prompt(none)►
◄get_name_image(none)►
◄set_namePrompt_Studio►
◄modeAuto►
◄raw_prompt►
◄model_backendGGUF / llama.cpp►
◄gguf_modelqwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q8_0.gguf►
◄mmproj_filemmproj-qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.gguf►
◄chat_handler▾►
◄text_encoder_namegemma4_e4b_it_fp8_scaled.safetensors►
◄clip_loader_typeAuto►
◄purpose▾►
◄vision_task▾►
◄model_format▾►
◄aesthetic▾►
◄extra_instructions►
◄system_prompt_override►
◄custom_instruction►
◄append_no_thinktrue►
◄n_gpu_layers-1►
◄n_ctx4096►
◄max_tokens1000►
◄temperature0.70►
◄top_p0.90►
◄repeat_penalty1.15►
◄seed0►
◄lock_infalse►
◄raw_prompt_input—►
Category ✨ TJ_Node/LLM
Inputs (30)
| Name | Type | Default | Description |
|---|---|---|---|
| get_name_prompt | COMBO | (none) | 1 options: (none) |
| get_name_image | COMBO | (none) | 1 options: (none) |
| set_name | STRING | Prompt_Studio | — |
| mode | COMBO | Auto | 3 options: Auto, Image to Prompt, Prompt Enhancer |
| raw_prompt | STRING | — | |
| model_backend | COMBO | GGUF / llama.cpp | 2 options: GGUF / llama.cpp, ComfyUI TextGenerate |
| gguf_model | COMBO | qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q8_0.gguf | 1 options: qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q8_0.gguf |
| mmproj_file | COMBO | mmproj-qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.gguf | 2 options: none, mmproj-qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.gguf |
| chat_handler | COMBO | 1 options: NO_VISION_HANDLERS_AVAILABLE | |
| text_encoder_name | COMBO | gemma4_e4b_it_fp8_scaled.safetensors | 1 options: gemma4_e4b_it_fp8_scaled.safetensors |
| clip_loader_type | COMBO | Auto | 29 options: Auto, krea2, stable_diffusion, stable_cascade, sd3, stable_audio, +23 |
| purpose | COMBO | 4 options: Image, Video, Edit (Inpainting/I2V), Custom | |
| vision_task | COMBO | 6 options: Caption (plain description), Caption + Format (apply model_format below), SD/Booru Tags, Pose & Anatomy Focus, Content Quality Check, Custom Instruction | |
| model_format | COMBO | 13 options: KREA2 (Prompt Enhance), Flux & Chroma (natural language), Z-Image & Lumina-2 (LLM text encoder), HiDream (hybrid prose + descriptors), SDXL (tags + weights), SD 1.5 (tags + weights), +7 | |
| aesthetic | COMBO | 22 options: None (no aesthetic injection), Photorealistic, Cinematic Film, Anime / Manga, Studio Ghibli, Pixar / 3D Animation, +16 | |
| extra_instructions | STRING | — | |
| system_prompt_override | STRING | — | |
| custom_instruction | STRING | — | |
| append_no_think | BOOLEAN | true | — |
| n_gpu_layers | INT | -1-1–999 | — |
| n_ctx | INT | 4096512–32768 | — |
| max_tokens | INT | 100050–4096 | — |
| temperature | FLOAT | 0.700–2 | — |
| top_p | FLOAT | 0.900–1 | — |
| repeat_penalty | FLOAT | 1.151–2 | — |
| seed | INT | 00–18446744073709550000 | — |
| lock_in | BOOLEAN | false | — |
| raw_prompt_inputopt | STRING | — | |
| imageopt | IMAGE | — | |
| clipopt | CLIP | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| prompt | STRING | — |
| thought_process | STRING | — |