Persona Image Generator
Generate a concrete image output (not just a prompt) for persona/avatar requests.
Workflow
- •Extract intent in one line: subject + style + mood + output usage.
- •Ask only missing essentials (max 3):
- •style (pixel-art, anime, cinematic, minimal, etc.)
- •framing (portrait, full-body, scene)
- •constraints (colors, outfit, symbols, no-go items)
- •Create one strong base prompt + 2 variations.
- •Invoke the configured image-generation capability (e.g., nano-banana-pro) and produce image outputs.
- •Return:
- •generated image(s)
- •short caption per image
- •reusable final prompt text
Prompt template
Use this structure:
<subject>, <visual style>, <camera/framing>, <lighting>, <palette>, <mood>, <environment>, high detail, clean composition, no text watermark
Defaults
- •If user says "draw your persona":
- •subject: friendly AI assistant mascot
- •style: modern pixel-art or clean illustration
- •framing: portrait (head + shoulders)
- •mood: warm, competent, calm
- •Generate 3 variants unless user asks for 1.
Quality checks
Before finalizing, verify:
- •subject matches requested identity
- •style is clearly visible
- •no unwanted text/watermarks
- •output is usable as avatar/banner (if requested)
Output format
Provide concise output:
- •
Variant A/B/C+ one-line style description - •Attach image(s)
- •
Prompt used:block for reuse/editing
Do not
- •Do not respond with prompt-only if user explicitly asked for generated images.
- •Do not over-question; proceed with sensible defaults when possible.
- •Do not include copyrighted character names unless user explicitly requests parody/fan style.