Can I create from text and from reference images?

Yes. The current Toono AI workspace supports both text-to-image and reference-guided image-to-image workflows for Flux 2 Pro. Give each reference a clear role when a product, material, palette, or composition needs to carry into the result.

How many reference images can I add?

The current workspace accepts up to eight reference images. More sources also increase the current credit calculation, so start with the smallest set that clearly defines the job.

Which resolutions, dimensions, and formats are available?

You can choose 0.5, 1, 2, or 4 MP; set custom width and height from 256 to 2048 pixels in multiples of 16; and export PNG, JPG, or WebP. The active controls and credit cost appear in the workspace.

Are text, colors, and reference fidelity guaranteed?

No. Review spelling, exact colors, product geometry, small parts, people, reflections, and every reference-dependent detail before publishing or handing the file to production.

Production image generation with deliberate canvas control

Build the image around the placement—not the other way round

Start from a prompt or up to eight reference images, then choose a 0.5–4 MP tier or set a custom canvas from 256 to 2048 pixels per side. Flux 2 Pro gives product and creative teams one workspace for composition, detail checks, and PNG, JPG, or WebP delivery.

Create with Flux 2 Pro

Open image generation with this model selected

See examples and review points

Reference images
0.5–4 MP
Custom size

A chartreuse modular chair moves from a controlled studio scene to material close-ups and channel-specific compositions while its silhouette and finishes stay reviewable.

Visual examples

Four checks between a promising direction and a usable file

A polished preview is only the start. Test whether references guide the intended composition, materials survive close inspection, the canvas suits the placement, and the product remains recognizable across a set.

Black Forest Labs

Guide the scene with explicit references

Assign product form, environment, palette, and composition to separate sources, then check scale, perspective, and contact in the combined result.

Black Forest Labs

Compose for the final canvas

Design the subject and negative space for the actual banner, card, or editorial crop instead of relying on a destructive crop after generation.

Black Forest Labs

Inspect material and construction detail

Zoom into textile weave, seams, tubular metal, fasteners, and reflections before treating the image as a catalog or campaign asset.

Black Forest Labs

Compare product consistency across scenes

Track silhouette, proportions, upholstery color, frame geometry, and distinctive details as the same chair moves into new environments.

Examples show production directions, not guaranteed replicas. Reference fidelity, color, text, small geometry, latency, credit use, and final detail vary with source material and settings. Review every output before publishing.

Practical workflow

Treat references and dimensions as production inputs

Decide what each source controls and where the final image will appear before spending credits on larger output.

01
Assign one job to every reference
Separate the product source from optional environment, material, palette, or composition references and remove conflicting directions.
02
Prove the direction at a practical size
Start with the canvas and resolution tier needed for comparison, then move to a larger output only after composition and subject detail hold up.
03
Inspect, export, and verify in context
Review at full size and inside the real placement, then choose PNG, JPG, or WebP according to the delivery workflow.

Flux 2 Pro workflow questions

Image inputs, output choices, and review points in the current Toono AI integration.

Compare another route before committing the brief

Use the open AI Image Generator for broader model choice, compare GPT Image 2 for text-led layouts, or try Nano Banana Pro for a different multi-reference workflow.