Multi-Engine Prompt Enhancer

Transform brief, simple ideas into highly detailed, optimized image generation prompts tailored perfectly for Midjourney, Stable Diffusion, DALL-E 3, Nano Banana, and text-to-image assistant models simultaneously.

Whether you are looking for an automated structural prompt template fallback or advanced LLM-powered context expansion via OpenRouter, this Actor handles the precise syntactic weightings, structural demands, and formatting flags that each image engine requires to deliver elite aesthetic results.

---

🚀 Features

• 🧠 Smart Context Expansion: Translates a simple subject, style, and mood into an immersive visual scene description.
• 🤖 OpenRouter LLM Integration: Connects seamlessly with cloud-hosted language models (such as Llama 3) to inject rich, intelligent prose modifiers.
• ⚙️ Multi-Engine Customization: Instead of a generic prompt output, it structures the results into a precise array of specialized prompts:
  • Midjourney: Appends modern technical aspect ratio configurations (--ar 16:9) and execution parameters (--v 6.0 --style raw).
  • Stable Diffusion (SDXL): Injects highly weighted, comma-separated quality tags (masterpiece, trending on artstation, vray render) to maximize fidelity.
  • DALL-E 3: Structures prompts into natural, organic conversational prose to bypass internal system revisions and match its semantic parser.
  • Nano Banana: Formats with fast-inference micro-rendering token prefixes.
  • GPT-2 Assistants: Provides instructional prompt frames designed for text transformers.

---

📥 Input Parameters

The Actor accepts the following properties via its user interface:

Field	Type	Required	Description
Main subject (subject)	String	Yes	The core entity or action you want to draw (e.g., "An armadillo wearing a trench coat").
Art style (style)	String	No	Predefined presets like watercolor, cinematic, pixel art, or your own custom artistic style.
Mood / atmosphere (mood)	String	No	Atmospheric parameters like mysterious, joyful, dark, or custom lighting cues.
Enhance with LLM (useLLM)	Boolean	No	When enabled, routes the base scene description through an LLM via OpenRouter for advanced creative prompt engineering.

---

📤 Output Format

The Actor returns a clean, structured JSON array containing objects mapped precisely to your target generators. This makes it incredibly easy to copy individual variants or stream the array directly into downstream automated imaging workflows.

Output JSON Example

[
  {
    "generator": "Midjourney",
    "optimizedPrompt": "A dramatic cinematic film still of a cyberpunk cat, shot on a 35mm anamorphic lens. Masterful chiaroscuro lighting, sharp focus on details... --ar 16:9 --v 6.0 --style raw",
    "notes": "Appended aspect ratio parameter and version configurations for Midjourney execution v6."
  },
  {
    "generator": "Stable Diffusion",
    "optimizedPrompt": "A dramatic cinematic film still of a cyberpunk cat... masterpiece, highly detailed, sharp focus, 8k resolution, trending on artstation",
    "notes": "Appended explicit comma-separated quality tags to trigger optimal high-fidelity weights in SDXL checkpoint pathways."
  }
]

### Setup Requirement

To use the LLM enhancement capability, you must provide your own API credentials:

Go to your Actor's Settings tab in the Apify Console.

Scroll down to Environment variables.

Create a variable with the key OPENROUTER_API_KEY and paste your secret token as the value.