Turn a text prompt into finished images in a single request. Pick a model with the
Learn more about Image from TextYour images show up here
Pick a style and a model — prices are per image, shown on the button.
Turn a text prompt into finished images in a single request. Pick a model with the model field — Google Imagen, OpenAI GPT Image, Black Forest Labs FLUX.2, Stability AI, Ideogram 4.0 or Leonardo — then add a style preset, optional negative prompts, aspect ratio, quality and how many images you want (1–4). Responses are synchronous — you get hosted image URLs back directly, with no polling.
Call List Image Models (media-image-models) for the live catalog: every model's provider, per-image price and supported aspect ratios.
Pricing is per image and depends on the model — from 3 credits (FLUX.2 Klein) to 20 credits (Ideogram 4.0 Quality). The default model flash (Imagen 3 Fast) is 4 credits/image.
The model field picks the provider: Imagen (flash, pro, imagen4-fast, imagen4), OpenAI (gpt-image-2), FLUX (flux-2-klein-4b/9b, flux-2-pro, flux-2-max), Stability (stable-core, stable-sd35-*, stable-ultra), Ideogram (ideogram-4-turbo/-4/-4-quality) or Leonardo (leonardo-lucid-*, leonardo-phoenix).
GET media-image-models (free) returns every model with its per-image price and aspect ratios — ideal for building a model picker.
From photorealistic to anime, cinematic, ukiyo-e, pixel-art, vaporwave, studio-ghibli, surreal and art-deco — call media-image-styles for the full list with descriptions.
aspect_ratio: 1:1 (default), 16:9, 9:16 or 4:3. quality: standard or hd (provider-dependent tiers).
Request 1–4 images per call. You're billed only for the images actually returned, at the selected model's per-image rate.
Images come back in the response as hosted URLs. No async job, no polling required.
Marketing & social posts
Generate hero images for blog posts, ads, and social campaigns from a brief prompt + style preset.
Product mockups
Quickly visualise concepts before commissioning real photography.
Editorial illustrations
Article headers, podcast cover art, newsletter visuals — pick a style preset matching your brand.
Game / app assets
Background art, character concepts, UI illustrations — multiple aspect ratios supported.
input
prompt + optional model (default: flash), style preset, negatives, aspect_ratio, quality, quantity
output
An array of generated images (hosted URLs) with id, url and index — plus the style, model, aspect_ratio, quality and generation time.
Each style preset adds descriptive text to your prompt to steer the look. Pick by intent:
Call media-image-models for the live list. Snapshot:
| Provider | Models | $/image | Notes |
|---|---|---|---|
| Imagen (Google) | flash (default), pro, imagen4-fast, imagen4 | 0.04 / 0.08 / 0.06 / 0.12 | Balanced default; imagen4 for top Imagen quality. |
| OpenAI | gpt-image-2 | 0.10 | Strong prompt following & text rendering. |
| FLUX (BFL) | flux-2-klein-4b, flux-2-klein-9b, flux-2-pro, flux-2-max | 0.03 / 0.03 / 0.06 / 0.14 | Klein = cheapest option overall. |
| Stability | stable-core, stable-sd35-flash/-medium/-turbo/-large, stable-ultra | 0.05–0.16 | SD 3.5 family + Ultra tier. |
| Ideogram | ideogram-4-turbo, ideogram-4, ideogram-4-quality | 0.06 / 0.12 / 0.20 | Excellent typography/text-in-image. |
| Leonardo | leonardo-lucid-origin, leonardo-lucid-realism, leonardo-phoenix | 0.08 / 0.08 / 0.10 | Stylised & realistic art looks. |
Rules of thumb: iterate on flux-2-klein-4b (3 credits) or flash (4 credits); text-heavy designs on Ideogram; finals on imagen4, flux-2-max or ideogram-4-quality.
Provide a comma-separated list of things to avoid. Common patterns:
low quality, blurry, watermark, text, distortedextra fingers, malformed hands, asymmetric eyessignature, frame, border, croppedNegatives reduce common artefacts but don't guarantee they're removed entirely.
promptrequiredstringText description of the image to generate.
stylestring · default: photorealisticStyle preset key (see media-image-styles for the full list of 25).
enum: photorealistic, anime, digital-art, oil-painting, watercolor, 3d-render, sketch, cinematic, cyberpunk, steampunk, minimalistic, vintage, futuristic, abstract, pop-art, impressionism, ukiyo-e, pixel-art, vaporwave, studio-ghibli, fantasy-art, gothic, surreal, film-noir, art-deco
negativesstringComma-separated list of things to avoid (e.g. 'low quality, blurry, watermark').
aspect_ratiostring · default: 1:1Output aspect ratio.
enum: 1:1, 16:9, 9:16, 4:3
qualitystring · default: standardOutput quality tier: 'low' | 'standard' | 'hd'. Interpretation is provider-dependent (OpenAI maps to low/medium/high; on Imagen, hd enables prompt enhancement).
enum: low, standard, hd
quantityinteger · default: 1Number of images to generate (clamped to 1-4).
modelstring · default: flashPicks the provider and model — use a model_id / call_as / alias from media-image-models. Imagen: 'flash' (default, 4 credits), 'pro' (8 credits), 'imagen4-fast' (6 credits), 'imagen4' (12 credits). OpenAI: 'gpt-image-2' (10 credits). FLUX: 'flux-2-klein-4b'/'−9b' (3 credits), 'flux-2-pro' (6 credits), 'flux-2-max' (14 credits). Stability: 'stable-core' (6 credits), 'stable-sd35-flash/-medium/-turbo/-large', 'stable-ultra' (5 credits-0.16). Ideogram: 'ideogram-4-turbo' (6 credits), 'ideogram-4' (12 credits), 'ideogram-4-quality' (20 credits). Leonardo: 'leonardo-lucid-origin'/'−realism' (8 credits), 'leonardo-phoenix' (10 credits).
data.imagesarrayArray of {id, url, index} per generated image.
data.images[].idstringPer-image identifier.
data.images[].urlstringHosted URL to the image file.
data.images[].indexinteger0-based index within this batch of images.
data.quantity_requestedintegerNumber of images you asked for (1–4).
data.quantity_generatedintegerNumber of images returned — this is what you're billed for.
data.modelstringResolved model id (e.g. flash, gpt-image-2, flux-2-pro).
data.providerstringProvider the model belongs to (imagen / openai / bfl / stability / ideogram / leonardo).
data.stylestringStyle preset applied.
data.aspect_ratiostringAspect ratio used.
data.qualitystringQuality tier used.
data.promptstringThe prompt you sent, returned for reference.
data.generation_time_secondsnumberHow long generation took, in seconds (informational).
cost.unitsnumberNumber of images billed (equals data.quantity_generated).
cost.unit_pricenumberThe selected model's per-image price — this is the rate you're billed at.
cost.tokensnumberTotal charged for this call (units × unit_price).
cost.balancenumberYour remaining credit balance after this call.
Per image, at the model's own rate. Listing models and style presets is free. media-image-models always returns the live prices.
| Service | Unit | Price |
|---|---|---|
| Imagen — flash (default) | item | 4 credits/image |
| Imagen — pro / imagen4-fast / imagen4 | item | 8 credits / 6 credits / 12 credits per image |
| OpenAI — gpt-image-2 | item | 10 credits/image |
| FLUX — flux-2-klein-4b/9b · flux-2-pro · flux-2-max | item | 3 credits · 6 credits · 14 credits per image |
| Stability — stable-core · sd35-flash/-medium/-turbo/-large · stable-ultra | item | 6 credits · 5 credits-0.13 · 16 credits per image |
| Ideogram — ideogram-4-turbo · ideogram-4 · ideogram-4-quality | item | 6 credits · 12 credits · 20 credits per image |
| Leonardo — lucid-origin / lucid-realism · phoenix | item | 8 credits · 10 credits per image |
| List Image Models / Image Styles | item | Free |
You're billed for the number of images actually returned (data.quantity_generated) × the selected model's per-image rate.
If you request 4 images and only 2 come back, you're billed for 2.
Cheapest: flux-2-klein (3 credits). Default: flash (4 credits). Premium: ideogram-4-quality (20 credits).