Which image models are behind the MCP server?

11, spanning every budget: flux-schnell ($0.003), z-image-turbo ($0.01), runway-gen4-image-turbo, nano-banana ($0.039), qwen-image-edit-plus and flux-2-pro ($0.03), seedream and recraft-v3 ($0.04), nano-banana-2 ($0.067), nano-banana-pro ($0.134), and gpt-image-2 ($0.128). One connection, no separate provider accounts.

How does an agent choose the right model?

Two ways: search_models matches models to a task description and price ceiling, and compare_models actually runs the prompt across candidates under a budget cap — returning each output with real cost and latency, so the choice is evidence, not vibes.

What does image generation cost at scale?

Per output, no subscription: a hundred drafts on z-image-turbo is $1.00; a hundred premium gpt-image-2 renders is $12.80. estimate_cost prices any batch upfront, and your monthly spending limit is a hard server-side cap.

What happens when a generation fails?

Automatic refund — you're only billed for outputs that exist. Requests are idempotent too, so an agent retrying a timed-out call can't double-charge you.

Can I use the generated images commercially?

Yes — outputs are yours to use in products, marketing, and client work, subject to each underlying model's usage terms. Everything is stored in your creation history, retrievable later via search_creations.

Which clients can call these image tools?

Any MCP client. claude.ai and ChatGPT connect over OAuth as custom connectors; Cursor, Claude Code, Codex, and VS Code connect with an nj_sk_ API key header. Same server, same tools, same pricing.

Create free key

Image generationKIYO

AI image generation your agent can call.

11 image models — FLUX.1 Schnell at $0.003 up to GPT Image 2 at $0.128 — behind one MCP server. Claude, ChatGPT, Cursor, or any MCP agent picks the model, generates, and pays per output.

Z-Image$0.01

Nano Banana 2$0.067

Seedream$0.04

GPT Image$0.128

Connect in 60 seconds

KIYO · one brief, four directions — real generate_image runs through NinjaChat.

Paste this. Watch it run.

First ask

A watercolor map of the Pacific Northwest for a hiking blog header — wide, muted greens.

generate_image

1 image on seedream · $0.04

Use case · brand stills

Four directions. One bottle that holds.

A sparkling-water brand, generated through generate_image — then the agent picks the keeper.

KIYO persimmon wall — **KIYO**02 Persimmon

Four directions. One bottle that holds the brand.

Models

11 engines. One tool call.

An editorial shot of a cobalt coat in white dunes — a real FLUX.2 Pro run

FLUX.2 ProFLUX.2 Pro

A flamenco dancer mid-turn under a single spotlight — a real Seedream run

Seedream 4.5Seedream

A red-sailed boat crossing a mirror-still turquoise fjord — a real Nano Banana 2 run

Google Nano Banana 2Nano Banana 2

A matte espresso machine in a shaft of morning light — a real Nano Banana Pro run

Google Nano Banana ProNB Pro

A red fox mid-leap over a frozen creek at sunrise — a real Nano Banana run

Google Nano BananaNano Banana

A hummingbird frozen mid-air at a fuchsia flower — a real Z-Image Turbo run

Z-Image TurboZ-Image

GPT Image 2's take on the tiny astronaut asleep on a crescent moon

GPT Image 2GPT Image

Tools

21 tools. These do the work here.

Make

generate_imageCreate images and download the full-resolution results.

Decide

compare_modelsTry one prompt across models with a spending limit.
search_modelsCheck model capabilities and prices.

Library

get_creationView a saved result and its generation details.

Money

estimate_costCheck the price before generating.

Also on the server

create_mediagenerate_speechedit_imagegenerate_videoimage_to_videoedit_videoextend_videoget_generation_statussearch_creationssave_creation_aslist_creative_entitiesupload_assetprepare_scrub_videopublish_pageget_account_balanceadd_credits

Setup

Connected in about a minute.

Pick your client — the steps for the others stay one tap away.

Claude4 steps

Open connector settings
In claude.ai, go to Settings → Connectors and choose “Add custom connector”. Your account and workspace must allow custom connectors.
2
Paste the NinjaChat endpoint
Name it NinjaChat and paste the server URL — nothing else to configure.
text
```
https://www.ninjachat.ai/api/mcp
```
Sign in when prompted
Sign in to NinjaChat and approve the connection. No API key needed. Enable NinjaChat in each chat or project where you want to use it.
Ask Claude to make something
Start a new chat and enable NinjaChat in the tools menu.

Cursor4 steps

Create an API key
Sign in at ninjachat.ai and create a key under Developers → Keys. Keys look like nj_sk_… and carry your spending limit.
2
Add the server to mcp.json
Create or edit ~/.cursor/mcp.json (or .cursor/mcp.json in a project) and add the NinjaChat server with your key.
~/.cursor/mcp.json · template
Replace nj_sk_YOUR_KEY with your API key before using this template.
```
{
  "mcpServers": {
    "ninjachat": {
      "url": "https://www.ninjachat.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer nj_sk_YOUR_KEY"
      }
    }
  }
}
```
Reload and verify
Open Cursor Settings → MCP and confirm ninjachat shows a green dot with the ninjachat tools. If not, reload the window.
Generate from the agent
Open a new chat in Agent mode, then send your prompt.