ComfyUI Node
Browser Prompt Enhancer
Unofficial integration that drives Brave Leo in a local browser to append real-time context to, or fully rewrite, your prompt. No API key required.
Browser Prompt Enhancer
- enhanced_prompt
- raw_response
- status
◄original_prompt►
◄moderewrite►
◄search_topic►
◄leo_personalityAct as an expert Prompt Engineer for image generation. Your task is to expand my basic concept into a highly detailed, "long-form" positive text-to-image prompt. Assume the model renders exactly what is described physically. Do not rely on it to infer missing details; every texture, light source, and material property must be explicitly stated in the positive prompt. Please follow this structure for the enhanced prompt: 1. Subject & Micro-Details: Define the main subject with intricate physical descriptions (e.g., pore structure, fabric weave, material imperfections, subsurface scattering). 2. Environment & Atmosphere: Describe the setting with depth, specific atmospheric particles (haze, dust, humidity), and background geometry. 3. Lighting Physics: Specify the light source (type, color temperature, angle), interaction with surfaces (specular highlights, ray-traced reflections, volumetric fog), and shadow density. 4. Composition & Camera: Define the lens focal length (e.g., 35mm, 85mm, macro), aperture (depth of field), camera angle, and framing rules. 5. Style & Render Quality: Explicitly state the rendering style (e.g., film grain, halftone, oil paint texture) and specific surface qualities (matte, glossy, rough, wet). Output ONLY the enhanced positive prompt, ready to be copied and pasted directly into an image model. Do not include any negative constraints, introductory text, or explanations.►
◄connection_modemanaged►
◄window_modeheadless►
◄timeout_seconds90►
◄brave_path►
◄debug_host127.0.0.1►
◄debug_port9222►
CategoryBrowser Prompt Enhancer
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| original_prompt | STRING | Your base prompt. Returned unchanged if Leo can't be reached. | |
| mode | COMBO | rewrite | append: search and add details to the end (keeps your intent). rewrite: search and rewrite the whole prompt for image generation. |
| search_topicopt | STRING | What Leo should search for. Empty = use the original prompt. | |
| leo_personalityopt | STRING | Act as an expert Prompt Engineer for image generation. Your task is to expand my basic concept into a highly detailed, "long-form" positive text-to-image prompt. Assume the model renders exactly what is described physically. Do not rely on it to infer missing details; every texture, light source, and material property must be explicitly stated in the positive prompt. Please follow this structure for the enhanced prompt: 1. Subject & Micro-Details: Define the main subject with intricate physical descriptions (e.g., pore structure, fabric weave, material imperfections, subsurface scattering). 2. Environment & Atmosphere: Describe the setting with depth, specific atmospheric particles (haze, dust, humidity), and background geometry. 3. Lighting Physics: Specify the light source (type, color temperature, angle), interaction with surfaces (specular highlights, ray-traced reflections, volumetric fog), and shadow density. 4. Composition & Camera: Define the lens focal length (e.g., 35mm, 85mm, macro), aperture (depth of field), camera angle, and framing rules. 5. Style & Render Quality: Explicitly state the rendering style (e.g., film grain, halftone, oil paint texture) and specific surface qualities (matte, glossy, rough, wet). Output ONLY the enhanced positive prompt, ready to be copied and pasted directly into an image model. Do not include any negative constraints, introductory text, or explanations. | A system-style instruction prepended to the request to Leo. |
| connection_modeopt | COMBO | managed | managed: launch a dedicated Brave profile (recommended). attach: connect to a Brave you started with --remote-debugging-port. |
| window_modeopt | COMBO | headless | headless: no window at all (recommended, fully invisible). hidden: real window pushed off-screen (may still show on Wayland/WSLg). visible: show the window (for debugging). managed mode only. |
| timeout_secondsopt | INT | 9010–300 | How long to wait for Leo's full response. |
| brave_pathopt | STRING | Override the Brave executable path. Empty = auto-detect. | |
| debug_hostopt | STRING | 127.0.0.1 | attach mode only: host of the running Brave's debug port. |
| debug_portopt | INT | 92221–65535 | attach mode only: the --remote-debugging-port you launched Brave with. |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| enhanced_prompt | STRING | — |
| raw_response | STRING | — |
| status | STRING | — |