Generative media, built into your product.

Generative media, built into your product.

Give your users AI-powered video, image and audio creation through one API. Powered by the world’s best models, and ready for production scale.

Start buildingTalk to sales

Trusted by

Ship the features your users want.

Add generative media to your product and give your users new ways to create.

Ad creative

Generate hundreds of video and image ad variants automatically.

Product visualization & video

Turn product photos into polished video, try-ons and lifestyle scenes — across entire catalogs.

Image & video editing

Clean up, restyle, upscale and reformat images and video automatically.

Personalized media

Generate video and images tailored to every audience and channel, automatically.

Create & social tools

Let your users create social posts and clips for every format and channel.

Real-time AI characters

Add conversational avatars for support, training and interactive experiences.

New capabilities start here.

As a frontier research lab, we're always pushing what's possible with generative media, so you can put the newest capabilities in front of your users first.

Video to Video

Aleph 2.0

Real-time Avatars

Characters

Text to Video

Gen-4.5

Text to Image

Gen-4 Image

Access the best models from across the industry

Explore all models

Text to Video

Seedance 2.0

Text to Image

GPT Image 2

Image to Image

Gemini 3 Pro

Text to Video

Veo 3.1

One API. Three ways to build.

Call models directly, build custom endpoints or use a pre-built endpoint that returns a finished result — all in one place.

Model access

Call models directly for video, image and audio, and get full control to build exactly what you want.

Workflows

Compose multi-step workflows that chain models and logic, and publish as custom endpoints.

Recipes

Get a finished result in one call from an endpoint we build and maintain — with our prompting and workflow expertise baked in.

Integrate in minutes.

Official SDKs for Python and Node, async task-based jobs and full type safety — call any model with a few lines of code.

NodePythoncURL

import RunwayML from '@runwayml/sdk';
const client = new RunwayML({ 
  apiKey: process.env.RUNWAYML_API_SECRET 
});
const task = await client.textToImage.create({
    model: "gemini_image3_pro",
    promptText: `A majestic snow leopard perched on a rocky cliff at sunset, golden light illuminating its fur, dramatic mountain landscape behind, hyper-detailed wildlife photography, 85mm lens, shallow depth of field`,
    ratio: "1344:768"
}).waitForTaskOutput();

A snow leopard on a rocky cliff at sunset, golden light on its fur, dramatic mountains behind.

google/gemini\_image3\_pro

Built for scale.

When your product runs on our API, your users are counting on it — so it's built for the reliability and scale they demand.

Talk to sales

99.9% uptime

SOC 2 Type II

No training on your data

IP indemnification

You own your outputs

Content moderation

Straight from the research lab.

Go deeper on the models and systems behind the API in technical write-ups from our research team.

Learn more about our research

Autoregressive-to-Diffusion Vision Language ModelsEfficient training of state-of-the-art diffusion vision language models.Read moreIntroducing Runway GWM-1A General World Model that simulates reality, built to be explored in real time.Read more

Build generative media into your platform.

Start buildingTalk to sales