Start Building →
Launching soon
Pixel Factory

Generate images and video.
On-demand, at scale.

Pixel Factory turns prompts into images and video through one simple API - Flux, SDXL, Stable Diffusion 3.5, and video models. Billed per output, with no GPUs to babysit.

API Docs
Per-image billing · text-to-image, image-to-image & video · seconds to first frame
POST · /v1/imagesrendering
prompt"a neon koi swimming through clouds"
flux.1
sdxl
sd3.5
video
i2i
upscale
Per-image
simple, predictable billing
Seconds
to first image
T2I · I2I · video
one unified API
US & EU
data residency
The API

One API. Pixels in seconds.

POST a prompt, get back image or video URLs. Sync for previews, async webhooks for batch jobs.

# pip install packet
import packet
 
job = packet.images.generate(
model="flux.1-dev",
prompt="a neon koi swimming through clouds",
size="1024x1024", n=4,
)
print(job.images[0].url)
Model catalog

The best open visual models.

Image and video models hosted and ready behind one endpoint. Indicative launch pricing per output.

ModelResolutionOutputFrom / image
Flux.1 [dev]Black Forest Labs
up to 2048²
Image
$0.025 Popular
Flux.1 [schnell]fast / 4-step
up to 1024²
Image
$0.003
SDXL 1.0Stability AI
up to 1024²
Image
$0.006
Stable Diffusion 3.5Large
up to 1024²
Image
$0.035
SVD-XTStable Video
24 fps
Video
$0.40 New
Real-ESRGANupscaler
4× / 8×
Upscale
$0.002
Video priced per second of output · image-to-image and inpainting at the same per-image rate.
Why Pixel Factory

Every pixel workflow, one endpoint.

Text-to-image

Flux, SDXL, and SD 3.5 behind a single API. Pick a model per request — no separate integrations.

Image-to-image & inpainting

Edit, extend, restyle, and inpaint from a reference image at the same per-image rate.

Video generation

Short clips from text or a reference image, returned via async webhook when the render completes.

ControlNet & LoRA

Pose, depth, and edge guidance, plus custom LoRA styles for on-brand, repeatable output.

Per-image billing

Pay per output, not per GPU-hour. No idle cost, no minimums, predictable per-asset economics.

Fast & async

Synchronous for quick previews, async with webhooks for high-volume batch generation.

Built for

Made for creative pipelines.

Marketing & ad creative

Generate on-brand variations at the speed of a campaign.

  • Batch variations
  • LoRA brand styles
  • Per-asset pricing

Product & ecommerce

Product imagery, backgrounds, and lifestyle shots on demand.

  • Image-to-image
  • Inpainting
  • Upscaling

Game & video

Concept art, 3D-asset references, and short video clips.

  • ControlNet
  • Video models
  • Async webhooks
How it works

Prompt in. Pixels out.

01

Get an API key

Create a key in the dashboard — no GPU provisioning, no render farm to manage.

02

POST a prompt

Send a prompt (and optional reference image) to /v1/images or /v1/video, choosing any model.

03

Receive output

Get image or video URLs back — synchronously for previews, or via webhook for batch.

FAQ

Pixel Factory, answered.

For anything not here, reach help@packet.ai.

What is Pixel Factory?
A managed image- and video-generation API. POST a prompt; packet.ai runs the diffusion models on GPUs and returns the output. No render farm to operate.
Which models can I use?
Flux.1, SDXL, Stable Diffusion 3.5, and video models like SVD-XT, plus an upscaler — selectable per request through one endpoint.
Can I do image-to-image and video?
Yes — text-to-image, image-to-image, inpainting, ControlNet guidance, and text/image-to-video, all through the same API.
How is it priced?
Per output: per image for stills, per second for video. No idle GPU cost and no minimums. Launch pricing starts at $0.003/image.
When does it launch?
Pixel Factory is in private preview. Join the waitlist for early access and launch-pricing credits.

Images and video,
without a render farm.

Pixel Factory is launching soon. Join the waitlist for early access and launch-pricing credits.

API Docs
Launching soon · per-image billing · images & video