run the world's best ai models

one line of code. instant api. pay per second.

build for free see pricing
code replicate.run("blaze/flux-pro")
text-to-image llms video audio
all systems operational

features

api

one-line api

Deploy any model with a single HTTP request. No infrastructure, no scaling headaches.

lightning

scale to zero

Pay only while your model runs. Instantly scales to zero when idle.

community

community models

Access over 10,000 community‑shared models or publish your own.

how it works

01

pick a model

Browse the gallery or search by modality. Every model has a ready‑to‑use endpoint.

02

run one line

Call the model with replicate.run().

03

ship it

Integrate the output into your app, workflow, or creative project.

what builders say

“Blaze cut our prototype time from weeks to hours. The API is absurdly simple.”
“I love that I only pay for actual compute. No more idle GPU bills.”
“The community models are a treasure trove. I found a diffusion model for medical imaging in minutes.”

pricing

explorer

$0

Free tier with limited monthly credits.

  • 100 predictions/month
  • Community models only
  • Standard priority
get started

team

custom

Volume discounts, SLAs, dedicated support.

  • Custom predictions/month
  • All models + private
  • Enterprise priority
  • Dedicated account manager
contact sales

frequently asked questions

imagine what you can build.