ComfyUI Node
LM Studio FLUX.2 Prompt Generator
Simple text-to-FLUX.2 prompt generator using LM Studio.
LM Studio FLUX.2 Prompt Generator
- enhanced_prompt
◄server_urlhttp://127.0.0.1:1234/v1/chat/completions►
◄modelqwen3.5-35b-a3b-uncensored-hauhaucs-aggressive►
◄base_promptA cinematic portrait of a person in a softly lit room.►
◄target_word_count170►
◄prompt_modephotoreal►
◄extra_requirements►
◄avoidtext, watermark, logo, extra fingers, deformed hands, low quality►
◄seed_modefixed►
◄seed42►
◄temperature0.30►
◄max_tokens4096►
CategoryLM Studio/FLUX.2
Inputs (11)
| Name | Type | Default | Description |
|---|---|---|---|
| server_url | STRING | http://127.0.0.1:1234/v1/chat/completions | LM Studio local chat completions endpoint. |
| model | STRING | qwen3.5-35b-a3b-uncensored-hauhaucs-aggressive | Exact LM Studio model ID from /v1/models. |
| base_prompt | STRING | A cinematic portrait of a person in a softly lit room. | Your short idea or draft prompt. The node will rewrite it as a FLUX.2 prompt. |
| target_word_count | INT | 17020–300 | Approximate prompt length in words. LLMs do not count exactly. |
| prompt_mode | COMBO | photoreal | Overall prompt style. |
| extra_requirements | STRING | Optional details that should be included or respected. | |
| avoid | STRING | text, watermark, logo, extra fingers, deformed hands, low quality | Things the generated prompt should avoid mentioning or should discourage. |
| seed_mode | COMBO | fixed | fixed uses the seed below; random creates a new seed on each run. |
| seed | INT | 420–2147483647 | Seed sent to LM Studio. |
| temperature | FLOAT | 0.300–1.5 | Creativity/randomness. 0.2-0.4 is recommended. |
| max_tokens | INT | 4096512–262144 | LM Studio output token budget. Increase this if prompts are incomplete or not clean. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| enhanced_prompt | STRING | — |