Actions11
- 💬 对话 / Chat Actions
- 🧠 深度思考 / Thinking Actions
- 🖼️ 图片 / Image Actions
- 🎬 视频 / Video Actions
- 📐 向量 / Embeddings Actions
- ✨ 提示词 / Prompt Actions
- 👗 虚拟试穿 / Virtual Try-On Actions
✨ 提示词 / Prompt → 优化 / Enhance
AI-generatedSummary
Enhances a prompt for image generation by adding details such as lighting, composition, background, and style based on the selected image category.
Inputs
- model (required) — The model to use for prompt enhancement. Options: Gemini 2.5 Flash (gemini-2.5-flash), GPT-4o (gpt-4o), or Custom (__custom).
- customModel — Custom model ID to use when 'Model' is set to Custom (__custom).
- prompt (required) — The original prompt to enhance.
- category (required) — Target image category that guides the enhancement. Options: product_photo, lifestyle, flat_lay, model_shot, detail_shot, packaging, social_media, banner.
- additionalOptions.style — Target visual style (e.g. minimal, luxury, vibrant) to incorporate into the enhanced prompt.
- additionalOptions.language — Output language for the enhanced prompt. Options: en (English), zh (Chinese). Default: en.
- additionalOptions.temperature — Controls randomness of the model output (0 = deterministic, 2 = more creative). Default: 0.7.
- additionalOptions.maxTokens — Maximum number of tokens allowed in the model response. Default: 2048.
- additionalOptions.simplify — Whether to return a simplified output containing only key fields. Default: true.
- additionalOptions.systemPromptOverride — Optional custom system prompt to replace the built‑in enhancement instructions. Leave empty to use the built‑in prompt.
- additionalOptions.extraBodyFields — Additional JSON fields to merge into the request body sent to the model (e.g. '{'top_p': 0.9}').
Output shape
object
When 'additionalOptions.simplify' is true (default), only the fields listed under 'Simplified output' are returned. When false, an extra 'raw_content' field is included with the model’s raw response text.
Examples
Example 1: Basic usage with default model and English output
Model: Gemini 2.5 Flash, Prompt: 'A red shoe', Category: product_photo, Language: en, Simplify: true
Example 2: Using a custom model and requesting Chinese output with full raw response
Model: __custom, Custom Model: 'my-model-v1', Prompt: 'A red shoe', Category: product_photo, Language: zh, Simplify: false