OpenAI-Compatible Endpoint

MuAPI exposes an OpenAI-compatible image generation API so any app that already speaks the OpenAI protocol works out of the box — no plugins, no code changes.

Field	Value
Base URL	`https://api.muapi.ai/v1`
Auth header	`Authorization: Bearer <your-api-key>`
Endpoints	`GET /v1/models` · `POST /v1/images/generations`

Available models

FLUX 1

Model ID	Description
`flux-schnell`	Fast generation (~5 s). Default.
`flux-dev`	Higher quality (~20 s).
`flux-kontext-dev`	FLUX Kontext T2I — dev.
`flux-kontext-pro`	FLUX Kontext T2I — pro.
`flux-kontext-max`	FLUX Kontext T2I — max.

FLUX 2

Model ID	Description
`flux-2-dev`	FLUX 2 Dev.
`flux-2-flex`	FLUX 2 Flex.
`flux-2-pro`	FLUX 2 Pro.
`flux-2-klein-4b`	FLUX 2 Klein 4B.
`flux-2-klein-4b-turbo`	FLUX 2 Klein 4B Turbo.
`flux-2-klein-9b`	FLUX 2 Klein 9B.
`flux-2-klein-9b-turbo`	FLUX 2 Klein 9B Turbo.

HiDream

Model ID	Description
`hidream-fast`	HiDream-I1 Fast.
`hidream-dev`	HiDream-I1 Dev.
`hidream-full`	HiDream-I1 Full — highest quality.

Seedream

Model ID	Description
`seedream`	ByteDance Seedream (v3).
`seedream-v4`	ByteDance Seedream v4.
`seedream-v4.5`	ByteDance Seedream v4.5.
`seedream-5.0`	Seedream 5.0.

Nano Banana

Model ID	Description
`nano-banana`	Nano Banana.
`nano-banana-pro`	Nano Banana Pro.
`nano-banana-2`	Nano Banana 2.

Kling Image

Model ID	Description
`kling-o1`	Kling O1 text-to-image.
`kling-o3`	Kling O3 text-to-image.

Hunyuan

Model ID	Description
`hunyuan-2.1`	Hunyuan Image 2.1.
`hunyuan-3.0`	Hunyuan Image 3.0.

Other models

Model ID	Description
`reve`	Reve text-to-image.
`ideogram-v3`	Ideogram v3.
`chroma`	Chroma image generation.
`sdxl`	Stable Diffusion XL.
`z-image`	Z-Image P.
`qwen`	Qwen text-to-image.
`vidu-q2`	Vidu Q2 text-to-image.
`grok-imagine`	Grok Imagine.
`grok-imagine-quality`	Grok Imagine (quality mode).
`wan2.1-t2i`	Wan 2.1 text-to-image.
`wan2.5-t2i`	Wan 2.5 text-to-image.
`wan2.6-t2i`	Wan 2.6 text-to-image.
`wan2.7-t2i`	Wan 2.7 text-to-image.
`wan2.7-t2i-pro`	Wan 2.7 text-to-image Pro.
`midjourney`	Midjourney v7 text-to-image.
`gpt4o`	GPT-4o text-to-image.
`gpt-image-2`	GPT Image 2 text-to-image.

DALL-E aliases

Model ID	Routes to
`dall-e-3`	`flux-schnell`
`dall-e-2`	`flux-schnell`

The dall-e-3 / dall-e-2 aliases let apps that hard-code DALL-E model names work without any configuration change.

Quick test

# List models
curl https://api.muapi.ai/v1/models \
  -H "Authorization: Bearer YOUR_API_KEY"

# Generate an image
curl https://api.muapi.ai/v1/images/generations \
  -X POST \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "flux-schnell",
    "prompt": "A golden retriever surfing on a neon wave, cinematic",
    "n": 1,
    "size": "1024x1024"
  }'

Response:

{
  "created": 1719411200,
  "data": [
    { "url": "https://cdn.muapi.ai/..." }
  ]
}

Supported sizes

`size`	Dimensions
`1024x1024`	Square (default)
`1792x1024`	Landscape
`1024x1792`	Portrait

Open WebUI

Go to Admin Panel → Settings → Images
Set Image Generation Engine to OpenAI
Set API Base URL to https://api.muapi.ai/v1
Set API Key to your MuAPI key
Set Image Model to flux-schnell (or flux-dev for higher quality)
Click Save — the 📷 image button in any chat thread is now powered by MuAPI

LM Studio

Open My Models → Image Models
Click Add Custom Endpoint
Set Base URL to https://api.muapi.ai/v1
Set API Key to your MuAPI key
Choose flux-schnell or flux-dev as the model

AnythingLLM

Go to Settings → AI Providers → Image Generation
Select OpenAI as the provider
Set Base URL to https://api.muapi.ai/v1
Set API Key to your MuAPI key
Set Model to flux-schnell

Any other OpenAI-compatible app

Point the app's OpenAI base URL to https://api.muapi.ai/v1 and provide your MuAPI key as the API key. If the app asks for an image model name, use any model ID from the table above — flux-schnell (fastest default), flux-dev, midjourney, gpt4o, nano-banana, seedream-5.0, and 40+ others. Apps that hard-code dall-e-3 or dall-e-2 also work without any changes.