Image to Prompt Generator

Upload a visual reference and turn its visible subject, composition, perspective, lighting, palette, materials, setting, and style into a reusable image generation prompt. Add an optional goal to preserve or change specific details.

How it works

Choose a reference image and optionally describe what the new prompt should preserve or change. Spira sends a browser optimized copy to a multimodal model and returns a primary prompt, useful alternatives and concise caveats. Analysis uses credits and requires sign-in.

1
Choose a reference image
Upload a PNG, JPG, WebP, GIF, BMP or AVIF image. The browser creates a smaller copy before it is sent for analysis.
2
Describe what matters
Optionally say which subject, composition, lighting or style should be preserved or changed in the generated prompt.
3
Generate and copy the prompt
Review the primary prompt, compare useful alternatives and copy the direction that best fits your next generation.

What you get

Simple tools, clear results, no unnecessary steps

More than a list of objects

Describe how the scene is framed, lit, colored, arranged, and styled so the prompt captures the visual relationship between its parts.

Prompt directions you can refine

Start with a detailed primary prompt and compare alternatives that vary useful visual choices without replacing the central subject.

Visible details organized for generation

The result brings together subject, framing, perspective, light, color, materials, environment and style in usable language.

Provider neutral wording

Use the prompt as a starting point in the image generator you prefer instead of receiving model specific parameters.

Clear limits on visual inference

The tool describes visible traits and uncertainty without claiming it recovered an original prompt or hidden context.

Use the reference as a starting point

The output describes visible traits but cannot recover the exact prompt, seed, model settings, or hidden intent behind an existing image.

Name the change you want

Ask to preserve the product, shift the lighting, simplify the background, or adapt the composition so the generated prompt supports your next task.

FAQ

Does the image stay in my browser?

No. The browser creates a smaller JPEG copy and sends it to the model for analysis. The page states this before generation.

Can it recover the original prompt?

No. It creates a new prompt from visible characteristics and does not know the original prompt, model, seed, or settings.

Which image formats can I upload?

You can choose PNG, JPG, WebP, GIF, BMP, or AVIF. The browser decodes and optimizes the image before analysis.

Turn a visual reference into a usable starting point

Generate a prompt, then refine the subject, composition and style for the image or video scene you actually need.