Image generation
29 image models, one API key, real prices
Nano Banana Pro, GPT Image 2, Seedream 5.0 Pro, FLUX 2 Pro, Gemini 3 Pro Image, Imagen 4 Ultra, Ideogram v3, Recraft v3, Grok Imagine and more — behind the same ak- key you already use for chat, with the exact USD cost shown before you generate.
Text-to-image and edit variants
29 image models today, from Google, OpenAI, ByteDance, Black Forest Labs, xAI and more. Full list at /models, live from GET /v1/models.
Up to 10 reference images
Drop in references on supporting models for edits, style transfer, and consistent characters. Adding a reference auto-switches the model to its edit variant — no separate model to pick.
Aspect, resolution, batch
1:1 through 21:9, 1K up to 4K where the model supports it, and multiple images per run. Controls only offer what the selected model actually accepts.
The price, before you spend it
The studio shows the exact USD cost — from roughly $0.01/image — computed by the same function that bills you. Failed generations are never charged.
Dashboard, API, or MCP
Use the visual studio at /images, call POST /v1/images/generations directly, or reach it from the apexapi-mcp server (generate_image tool) and the n8n community node.
Guardrails and stable hosting
Content guardrails and per-account rate/concurrency limits apply on every entry point. Results are re-hosted on our own CDN — stable URLs, not expiring provider links.
The flagships
29 image models total. These are the ones people ask for by name.
Call it directly
Synchronous, OpenAI-compatible: POST /v1/images/generations with a model and a prompt.
curl https://api.apexapi.dev/v1/images/generations \
-H "Authorization: Bearer ak-your-key" \
-H "Content-Type: application/json" \
-d '{
"model": "google/nano-banana-pro",
"prompt": "a golden retriever puppy in a knitted sweater, cozy morning light"
}'Image generation FAQ
How many image models are available?
29 image models in the live catalog today, out of 139 models total across the API. The full list with pricing is at apexapi.dev/models, and it's always current — the same data backs GET /v1/models.
How much does an image cost?
It depends on the model and settings, starting from roughly $0.01/image. The composer at /images shows the exact USD price for your current selection before you click generate, using the same pricing function that bills you. If a generation fails, you are not charged.
Can I use my own images as references?
Yes, on supporting models — up to 10 reference images, for edits, style transfer, and consistent characters. Adding a reference automatically switches the model to its edit variant, so you don't have to pick a separate model yourself.
Do I need a separate account or key for images?
No. It's the same ak- key and the same prepaid credit balance you already use for chat models. Credits are real US dollars, not an opaque token unit.
Is there an API, not just the dashboard?
Yes. POST /v1/images/generations takes an OpenAI-compatible shape (model, prompt) and returns synchronously. It's also exposed as the generate_image tool on the apexapi-mcp MCP server and through the n8n community node.
What happens to the images afterward?
Results are re-hosted on our own CDN, so you get stable URLs instead of a provider link that expires in a few hours. Everything you generate stays in your gallery, with a filter for whether it came from the dashboard, the API, or MCP.
Generate your first image
Open the studio, pick a model, and see the price before you spend it — on ApexApi.