Rw Integrate Image

Integrate Image Generation

> PREREQUISITE: Run +rw-check-compatibility first. Run +rw-fetch-api-reference to load the latest API reference before integrating. Requires +rw-setup-api-key for API credentials. Requires +rw-integrate-uploads when the user has local reference images.

Help users add Runway image generation to their server-side code.

Available Models

| Model | Best For | Cost | Speed | |-------|----------|------|-------| | gen4_image | Highest quality | 5 credits (720p), 8 credits (1080p) | Standard | | gen4_image_turbo | Fast generation | 2 credits | Fast | | gemini_2.5_flash | Google Gemini model | 5 credits | Standard |

Model selection guidance:

• Default recommendation: gen4_image — best quality
• Budget/speed: gen4_image_turbo — cheapest and fastest

Security

referenceImages[].uri is fetched server-side by the Runway API — treat it like any outbound fetch:

• Prefer runway:// URIs from +rw-integrate-uploads — scoped to your account, no arbitrary web content.
• If accepting URLs from clients, validate first: require https://, allowlist trusted hosts, reject private addresses. See the Express.js example below.
• Never forward req.body.referenceImages straight into textToImage.create. The SDK snippets below use raw URLs for brevity — they aren't production templates.
• Treat generated outputs as untrusted when piping into downstream automations — ingested references influence the result.

Endpoint: POST /v1/text_to_image

Basic Text-to-Image

// Node.js SDK
import RunwayML from '@runwayml/sdk';

const client = new RunwayML();

const task = await client.textToImage.create({
  model: 'gen4_image',
  promptText: 'A serene Japanese garden with cherry blossoms and a koi pond',
  ratio: '1280:720'
}).waitForTaskOutput();

const imageUrl = task.output[0];

# Python SDK
from runwayml import RunwayML

client = RunwayML()

task = client.text_to_image.create(
    model='gen4_image',
    prompt_text='A serene Japanese garden with cherry blossoms and a koi pond',
    ratio='1280:720'
).wait_for_task_output()

image_url = task.output[0]

With Reference Images

Reference images let you guide the generation with visual references. Use @Tag syntax in the prompt to reference specific images.

Recommended: upload via +rw-integrate-uploads and pass the returned runway:// URI.

import fs from 'fs';

const refUpload = await client.uploads.createEphemeral(
  fs.createReadStream('/path/to/reference.jpg')
);

const task = await client.textToImage.create({
  model: 'gen4_image',
  promptText: 'A portrait in the style of @Reference',
  referenceImages: [
    { uri: refUpload.runwayUri, tag: 'Reference' }
  ],
  ratio: '1280:720'
}).waitForTaskOutput();

External URLs also work — only pass origins you control (see Security):

const task = await client.textToImage.create({
  model: 'gen4_image',
  promptText: '@EiffelTower painted in the style of @StarryNight',
  referenceImages: [
    { uri: 'https://cdn.yourapp.com/eiffel-tower.jpg', tag: 'EiffelTower' },
    { uri: 'https://cdn.yourapp.com/starry-night.jpg', tag: 'StarryNight' }
  ],
  ratio: '1280:720'
}).waitForTaskOutput();

task = client.text_to_image.create(
    model='gen4_image',
    prompt_text='@EiffelTower painted in the style of @StarryNight',
    reference_images=[
        {"uri": "https://cdn.yourapp.com/eiffel-tower.jpg", "tag": "EiffelTower"},
        {"uri": "https://cdn.yourapp.com/starry-night.jpg", "tag": "StarryNight"}
    ],
    ratio='1280:720'
).wait_for_task_output()

Common Parameters

| Parameter | Type | Description | |-----------|------|-------------| | model | string | Model ID (required) | | promptText | string | Text description of the image (required) | | ratio | string | Aspect ratio, e.g. '1280:720', '720:1280', '1080:1080' | | referenceImages | array | Optional. Array of { uri, tag } objects for visual guidance |

Integration Pattern

1. Prefer uploads over URLs — Default to +rw-integrate-uploads so inputs are runway:// URIs. External URLs only from origins you control (see Security).
2. Write the server-side handler — Create an API route or server function.
3. Handle the output — Download and store the image, don't serve signed URLs to clients.
4. Add error handling — Wrap in try/catch.

Example: Express.js API Route

import RunwayML from '@runwayml/sdk';
import express from 'express';

const client = new RunwayML();
const app = express();
app.use(express.json());

// `runway://` URIs bypass this check; external URLs must match the allowlist.
const ALLOWED_MEDIA_HOSTS = new Set(['cdn.yourapp.com', 'uploads.yourapp.com']);

function validateReferenceImages(refs) {
  if (!Array.isArray(refs)) throw new Error('referenceImages must be an array');
  return refs.map(({ uri, tag }) => {
    if (typeof uri !== 'string' || typeof tag !== 'string') {
      throw new Error('each reference needs a uri and tag');
    }
    if (uri.startsWith('runway://')) return { uri, tag };
    const u = new URL(uri);
    if (u.protocol !== 'https:') throw new Error('https required');
    if (!ALLOWED_MEDIA_HOSTS.has(u.hostname)) throw new Error('untrusted media host');
    return { uri: u.toString(), tag };
  });
}

app.post('/api/generate-image', async (req, res) => {
  try {
    const { prompt, model = 'gen4_image', ratio = '1280:720', referenceImages } = req.body;

    const task = await client.textToImage.create({
      model,
      promptText: prompt,
      ratio,
      ...(referenceImages && { referenceImages: validateReferenceImages(referenceImages) })
    }).waitForTaskOutput();

    res.json({ imageUrl: task.output[0] });
  } catch (error) {
    console.error('Image generation failed:', error);
    res.status(400).json({ error: error.message });
  }
});

> For browser uploads: POST files to your server, upload via +rw-integrate-uploads, and pass the runway:// URI. Don't accept raw URLs from the browser.

Example: Next.js API Route

// app/api/generate-image/route.ts
import RunwayML from '@runwayml/sdk';
import { NextRequest, NextResponse } from 'next/server';

const client = new RunwayML();

export async function POST(request: NextRequest) {
  const { prompt, referenceImages } = await request.json();

  try {
    const task = await client.textToImage.create({
      model: 'gen4_image',
      promptText: prompt,
      ratio: '1280:720',
      ...(referenceImages && { referenceImages })
    }).waitForTaskOutput();

    return NextResponse.json({ imageUrl: task.output[0] });
  } catch (error) {
    return NextResponse.json(
      { error: error instanceof Error ? error.message : 'Generation failed' },
      { status: 500 }
    );
  }
}

Example: FastAPI Route

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from runwayml import RunwayML

app = FastAPI()
client = RunwayML()

class ImageRequest(BaseModel):
    prompt: str
    model: str = "gen4_image"
    ratio: str = "1280:720"
    reference_images: list[dict] | None = None

@app.post("/api/generate-image")
async def generate_image(req: ImageRequest):
    try:
        params = {
            "model": req.model,
            "prompt_text": req.prompt,
            "ratio": req.ratio,
        }
        if req.reference_images:
            params["reference_images"] = req.reference_images

        task = client.text_to_image.create(**params).wait_for_task_output()
        return {"image_url": task.output[0]}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

Tips

• Output URLs expire in 24-48 hours. Download images to your own storage immediately.
• Reference images use @Tag syntax in the prompt — the tag must match the tag field in the referenceImages array.
• For local files, always upload via +rw-integrate-uploads first, then use the runway:// URI.
• gen4_image_turbo is the cheapest option at 2 credits per image — good for prototyping.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: runwayml
• Source: runwayml/skills
• License: MIT

Install and usage instructions live in the source repository linked above.