The world’s fastest generative media

Add generative media into your application with the flexibility to use any SD model, LoRA, ControlNet, or IP-Adapter. No hardware, infra or AI expertise required – everything is a single API call away.

img
img
img
img
img
0.4s
FLUX Inference Time
0.3s
SD1.5 Inference Time
0.6s
SDXL Inference Time
2.5B
Images Generated

Run any workflow

Text to Image

Generate images from a text prompt. Use any model or bring your own and get complete freedom over all parameters, including ControlNet and LoRA.
img

Image to Image

Add a seed image alongside your text prompt to guide the AI image generation, and set how similar you'd like the results to be.
img

Background removal

Remove the background from any image to get a clean image and isolate people or objects.
img

Upscaling

Increase the resolution of an image by 2x, 3x or 4x, to obtain a higher quality image with more detail.
img

Inpainting

Transform any section of an image with just one API call, for both uploaded and AI-generated images.
img

Outpainting

Add new areas to an image by expanding it in any direction. Perfect for achieving a zoom out effect.
img

Text to VideoCOMING SOON

Coming soon - Transform text into relevant, high-quality videos.
img

Pick from 194,136 models

Experiment with any models — everything from FLUX to SD1.5, SDXL/Turbo, SD3, LoRAs, ControlNet and embeddings. Add any public or private model for complete flexibility.
img
ReV Animated
img
DreamShaper
img
Juggernaut XL
img
RealVision
LoRA
img
Bathroom - JJ’s ...
img
epiCRealism
img
Ghostmix
img
Absolute Reality
img
TrueSketch
LoRA
img
Detail Tweaker
img
Tales of Digital Life
img
MOHAWK
img
Unstable Diffusers
LoRA
img
Picture Book
img
Boltning
img
FLUX
img
ReV Animated
img
DreamShaper
img
Juggernaut XL
img
RealVision
LoRA
img
Bathroom - JJ’s ...
img
epiCRealism
img
Ghostmix
img
Absolute Reality
img
TrueSketch
LoRA
img
Detail Tweaker
img
Tales of Digital Life
img
MOHAWK
img
Unstable Diffusers
LoRA
img
Picture Book
img
Boltning
img
FLUX
img
ReV Animated
img
DreamShaper
img
Juggernaut XL
img
RealVision
LoRA
img
Bathroom - JJ’s ...
img
epiCRealism
img
Ghostmix
img
Absolute Reality
img
TrueSketch
LoRA
img
Detail Tweaker
img
Tales of Digital Life
img
MOHAWK
img
Unstable Diffusers
LoRA
img
Picture Book
img
Boltning
img
FLUX
img
ToonYou
img
LEOSAM's HelloWorld
img
Robmix
img
Socarealism
img
Photopedia
img
Opendalle
img
RealCartoon3D
img
dvArch
img
Crystal Clear One
img
Cinematix
LoRA img
Cyberpunk
img
Art Universe
img
Realistic Mix XL
img
CFXL
img
Easy Illustrator
img
ToonYou
img
LEOSAM's HelloWorld
img
Robmix
img
Socarealism
img
Photopedia
img
Opendalle
img
RealCartoon3D
img
dvArch
img
Crystal Clear One
img
Cinematix
LoRA img
Cyberpunk
img
Art Universe
img
Realistic Mix XL
img
CFXL
img
Easy Illustrator
img
ToonYou
img
LEOSAM's HelloWorld
img
Robmix
img
Socarealism
img
Photopedia
img
Opendalle
img
RealCartoon3D
img
dvArch
img
Crystal Clear One
img
Cinematix
LoRA img
Cyberpunk
img
Art Universe
img
Realistic Mix XL
img
CFXL
img
Easy Illustrator

Mix in any of the latest technologies

No curated model libraries or rigid options. We offer a truly flexible API, allowing you to use any cocktail of models, including your own. And what’s more, our proprietary Sonic Inference Engine® means we can do this in milliseconds.
  • Image control img
    • ControlNet
    • IP-AdapterPRIVATE BETA
    • Instant IDPRIVATE BETA
  • Model adaptersimg
    • LoRA
    • LCM-LoRAPRIVATE BETA
    • LyCORISPRIVATE BETA
  • Textual InversionPRIVATE BETA
  • AnimateDiffPRIVATE BETA
img

ControlNet

Precision control from a single input. With built-in customization and the ability to integrate your own, we've got you covered.
To request Beta access join us on Discord
img

IP-Adapter

Inspire a new image with the content of another image, helping you to create the perfect image variations for any scenario.
To request Beta access join us on Discord
img

Instant ID

InstantID is a cutting-edge method for creating ID-Preserving images using a single seed image. Ready to use out-of-the-box.
To request Beta access join us on Discord
img

LoRA

LoRAs enhance customization, acting as modifiers on a base checkpoint model. Integrate any CivitAI LoRA or your own with just an API call.
To request Beta access join us on Discord
img

LCM-LoRA

Accelerate high-quality image creation and reduce processing steps with LCM-enhanced Latent Diffusion Models.
To request Beta access join us on Discord
img

LyCORIS

Modify pre-trained models and inject or integrate specific characters into images, such as celebrities, fictional characters, cartoon characters, etc.
To request Beta access join us on Discord
img

Textual Inversion

Add small models that customize the output of your Stable Diffusion images, augmenting your SD output with specialized subjects and artistic styles.
To request Beta access join us on Discord

AnimateDiff

Transform text and static images into fluid, animated videos, making sophisticated animations readily available in any application.
To request Beta access join us on Discord

Total control

Take full command of your AI models with generation parameters. Fine-tune and tweak every detail to match your vision perfectly.
# of steps
Clip Skip
Seed
CFG Scale
Schedulers
Custom VAE
Prompt Weights
img
img
Quick tweaks or detailed perfection, adjust the steps to get the speed and quality you need.
img
Inference Steps
5
img
Inference Steps
10
img
Inference Steps
15
Balance focus and detail by managing how much of the prompt the model emphasizes.
img
CLIP Skip
3
img
CLIP Skip
2
img
CLIP Skip
1
Ensure consistent results with the same initial random noise, perfect for iterative testing.
img
Seed
Fixed seed
img
Seed
Fixed seed + prompt change purple to golden
img
Seed
Random seed + Prompt change
Set the creativity level. Lower values for more freedom, higher for strict adherence to your prompt.
img
CFG scale
4
img
CFG scale
8
img
CFG scale
12
Choose the right sampler for your project, from Euler to DDIM, and more.
img
Scheduler
Euler
img
Scheduler
DPM 2++ Karras
img
Scheduler
DDIM
Enhance image quality and specificity with custom Variational Autoencoders.
img
Custom VAE
Photon + vae-ft-mse-840000-ema-pruned
img
Custom VAE
Photon + clear VAE
img
Custom VAE
Photon + nocturnal VAE
Prioritize key elements in complex prompts by assigning weights.
img
Prompt Weights
Windy day:0.5
img
Prompt Weights
Windy day:1
img
Prompt Weights
Windy day:1.8

Platform extras

img

Prompt enhancer

Give your end users a head start with prompting. Magic Prompt suggests additional keywords or entirely new prompts based on a given text input.
img

Image to Prompt

Extract keywords from any image for precise prompting, and pair with Image to Image inference to create your desired image or style.
img

NSFW filter

Keep your generation results and community family-friendly. Easily toggle the filter on or off to ensure suitable images are created for your audience.
img

Image caching

Experience swift loading times with a 24-hour high-performance caching layer. Long-term storage options also available.
img

Pipeline implementations

Enhance your AI workflows with our help: achieve efficient data ingestion, automated processing, and scalable performance.
img

Cloud transfer

Opt for auto-sync of generated images to your cloud storage for easy integration and seamless backup or feature development.
img

Image storageADD-ON

Turn on our extended storage solutions, ensuring preserved and always available visuals, delivered via Cloudflare CDN.
img

Model storageADD-ON

Speak to us to have your models uploaded and cached, ensuring they are always available with the lowest possible latencies.