Skip to main content
Back

Runway Recipes Let Developers Ship AI Video Ads With One API Call

Runway's new API Recipes package its own prompting and workflow expertise into single-call endpoints for production-ready generative media features like product ads and video product swaps.

  • Runway launched API Recipes: pre-built, single-call endpoints with Runway's own prompting and workflow expertise packaged in.
  • Launch use cases: create product video ads from an image, and swap a product in an existing video clip.
  • No prompt engineering needed: Runway handles model selection, chaining, and prompting internally -- developers just supply inputs.
  • Powered by Gen-4.5 and Aleph 2.0: Aleph 2.0 enables localized product swaps; Gen-4.5 drives high-fidelity image-to-video ad generation.
  • Credit-based pricing: API credits cost $0.01 each; Recipe-specific pricing depends on underlying generation steps.
  • Available now via the Runway Developer Portal with Build and Enterprise tiers.

Building a generative media feature into your product has always meant two jobs: picking the right model, and then figuring out how to prompt it well enough to get consistent, polished results. Runway just shipped a feature that tries to collapse both into a single API call. Runway API Recipes are pre-built endpoints where Runway's own prompting and workflow expertise is baked in, so you get production-ready output without having to engineer or maintain the pipeline yourself.

What Recipes actually are

The developer portal now offers two distinct paths: build your own endpoints with Workflows, or use pre-built Recipes for specific outcomes. Workflows are the existing flexible primitive, where you chain together Runway's models yourself. Recipes are the opposite: opinionated, outcome-focused endpoints where Runway has already done the prompt engineering and model-chaining work for you.

The launch use cases are squarely aimed at commercial media production:

  • Product ad creation -- generate a polished video ad from a single product image
  • Product swap in video -- replace a product appearing in an existing video clip with a different one

Both are available today via the Runway API. You build your own endpoints with Workflows, or use the pre-built Recipes for specific outcomes -- try in the playground, ship with one API.

The problem this solves

Getting a generative model to produce something that looks good once is easy. Getting it to produce something that looks consistently good, at scale, across thousands of product images from different customers -- that's a much harder prompt engineering and orchestration problem. Most teams end up spending weeks tuning prompts, chaining models, and handling edge cases before they can ship anything.

If you are building an AI media API, the lesson is to separate the user intent layer from the model execution layer. Recipes are Runway's answer to that: they own the execution layer so you only have to supply the intent (and the image).

Publishing Workflows as API endpoints lets you build custom generative pipelines in the web app and deploy them with a single click -- no infrastructure setup or model orchestration required. Recipes take this further by removing the pipeline-building step entirely for common use cases.

What's under the hood

Runway hasn't published the internal architecture of each Recipe, but the underlying model stack is well-documented. Runway Aleph 2.0 is Runway's upgraded flagship video editing model for transforming existing footage while keeping the rest of the clip stable. It is built for precise localized edits, single-frame visual guidance, and consistent edits across cuts and scene changes, making it useful for product swaps, packaging updates, background changes, relighting, and scene cleanup. That makes it the obvious backbone for the product swap Recipe.

For image-to-video ad generation, Gen-4.5 is an AI video generation model that creates short video clips from text prompts or static images with high visual fidelity and smooth motion. It supports both text-to-video and image-to-video generation with a range of aspect ratios and clip durations. Gen-4.5 emphasizes realistic motion, strong prompt adherence, and controllable composition, making it suitable for cinematic sequences and creative video workflows.

The Runway API has been steadily expanding its model roster. Recent additions include:

  • Seedance 2.0, which generates high-quality videos from text, images, or video with support for keyframe control, reference images, reference videos, and generated audio.
  • OpenAI's GPT Image 2, supporting up to 16 reference images with optional prompt-referenceable tags, 30 width:height ratios across 1K, 2K, and 4K tiers.
  • Gemini 3 Pro Image, which generates images with up to 5,500-character prompts and up to 14 reference images, at 1K, 2K, or 4K resolution across 10 aspect ratios.

Recipes can draw on any of these models internally, and Runway can swap or upgrade the underlying model without breaking your integration -- since you're calling the Recipe endpoint, not a specific model identifier.

The real shift: intent over model selection

Runway's 2026 product direction says something important about the AI video market: the future is not a single model. It is a model marketplace wrapped in a creative workflow. Recipes are the sharpest expression of that strategy so far. Instead of asking developers to know which model handles product preservation best, or how to prompt for consistent lighting across a product swap, Runway absorbs that complexity.

AI video tools are becoming routers. In early AI video products, the model was the product -- you opened a tool because it had one special model. That breaks down once every model is good at something different. Users do not want to memorize all of that. They want the right output. The platform that wins is the one that hides model complexity when it should and exposes model control when experts need it.

Pricing and access

Each generation run via the Runway API costs credits. Credits can be purchased for $0.01 per credit in the developer portal for an organization. Recipe pricing hasn't been separately published -- it will depend on the underlying generation steps involved, which vary by Recipe. The API has two tiers: a Build plan for individuals and teams building internal products, and an Enterprise plan for larger organizations. The Runway API has been used by the world's largest consumer technology companies to reliably generate millions of videos for their users.

To get started, head to the Runway Developer Portal, grab an API key, and look for the Recipes section in the playground. The official documentation covers the full endpoint reference. A quick example of how you'd call a Recipe follows the same SDK pattern as any other Runway endpoint:

import RunwayML from '@runwayml/sdk';
const client = new RunwayML({ apiKey: process.env.RUNWAYML_API_SECRET });
// Call a Recipe endpoint -- Runway handles prompting and model routing
const task = await client.recipes.productAd.create({
  productImage: 'https://your-cdn.com/product.jpg',
  outputRatio: '16:9'
}).waitForTaskOutput();
console.log(task.output.videoUrl);

Who should care

Recipes are most useful if you're building a platform that needs generative media as a feature -- an e-commerce tool, a marketing automation product, a creative SaaS -- and you don't want to hire a prompt engineer or maintain a multi-model pipeline. The tradeoff is control: you get polished, consistent output, but you can't tune the internals the way you could with a raw Workflow.

If you need full control over every generation parameter, Workflows remain the right path. But for teams that just want a reliable "product image in, ad video out" endpoint they can ship to customers today, Recipes are a meaningful shortcut. The broader signal is that Runway is betting its API moat on workflow expertise, not just model quality -- and that's a bet worth watching as the generative video market matures.