Skip to content
ArcBrush ArcBrush Docs

AI Text to Image

Node icon
AI

Generates images from text prompts using AI, with five models from fast drafts to designer-grade output with readable in-image text.

AI Text to Image creates images from a text description. Five models cover different strengths in speed, cost, and specialty: FLUX Fast (quick iterations), FLUX.2 Pro (professional output), Nano Banana 2 (best prompt understanding, web search), FLUX.2 Max (highest fidelity), and GPT Image 2 (designer-grade output with the best in-image text rendering of any model, ideal for signage, UI mockups, posters, and multilingual labels).

The output is a standard image that wires into any other node: feed it through Palette Remap, add effects, composite it, or export directly. AI is just another node in the pipeline.

Parameters

Section titled “Parameters”

Pins

ImageImage
Output

Prompt

PromptText

Text prompt describing the image

Negative PromptText

What to avoid in the image (model-specific support)

Model

ModelChoice
Default: 0 Range: FLUX Fast / FLUX.2 Pro / Nano Banana 2 / FLUX.2 Max / GPT Image 2

AI model

Generation

SeedInteger
Default: -1 Range: -1–999999

Generation seed; -1 = random each run

Output Width (px)Integer
Default: 1024 Range: 64–2048
Model = FLUX Fast, FLUX.2 Pro or FLUX.2 Max

Output width

Output Height (px)Integer
Default: 1024 Range: 64–2048
Model = FLUX Fast, FLUX.2 Pro or FLUX.2 Max

Output height

StepsInteger
Default: 4 Range: 1–50
Model = FLUX Fast

Inference steps. Max depends on the selected model

Guidance ScaleNumber
Default: 3.5 Range: 1.0–20.0
Model = FLUX Fast

Prompt adherence strength

Aspect RatioChoice
Default: 6 Range: Auto / 21:9 / 16:9 / 3:2 / 4:3 / 5:4 / 1:1 / 4:5 / 3:4 / 2:3 / 9:16
Model = Nano Banana 2

Output aspect ratio

ResolutionChoice
Default: 1 Range: 0.5K / 1K / 2K / 4K
Model = Nano Banana 2

Output resolution (cost scales)

Web SearchToggle
Default: false
Model = Nano Banana 2

Enable web search augmentation (+1 credit)

Output SizeChoice
Default: 1 Range: 1024x768 (4:3) / 1024x1024 (1:1) / 1024x1536 (2:3) / 1920x1080 (16:9) / 2560x1440 (16:9 QHD) / 3840x2160 (16:9 4K)
Model = GPT Image 2

GPT Image 2 output size preset

QualityChoice
Default: 1 Range: Low / Medium / High
Model = GPT Image 2

Generation quality. Low/Medium cost less

Usage Tips

Section titled “Usage Tips”
  • Start with FLUX Fast for prompt iteration, then switch to FLUX.2 Pro or FLUX.2 Max for final output.
  • Reach for GPT Image 2 when the image needs readable text inside it (posters, UI mockups, labels) or when a complex multi-clause prompt has to be followed exactly.
  • GPT Image 2 prices a 6×3 matrix of output size × quality (Low / Medium / High). The default Medium is the sweet spot; the live cost label above Generate updates as you change either.
  • Use a negative prompt to exclude unwanted elements (FLUX Fast only).
  • Wire the output into Remove BG to get a clean cutout on transparent background.
  • Set a specific seed to reproduce a result you like, then tweak the prompt while keeping the same composition. (Seed is unsupported on GPT Image 2 and is hidden for that model.)
Section titled “Related Nodes”