Migrate from Runway ML to Railwail

TL;DR — Switch in Under 15 Minutes

• Drop Runway's task-based polling pattern — Railwail's POST /v1/videos blocks until complete
• Runway Gen-3 Alpha Turbo maps to gen-3-alpha-turbo on Railwail (mirrored)
• Plus Veo-3, WAN-2.5, AnimateDiff and Stable Video as alternatives
• EU hosting, EUR billing
• Same key unlocks the rest of the AI stack (text, image, audio)

Why Move Off Runway ML Direct?

Runway pioneered consumer-facing AI video and remains a strong choice for cinematic Gen-3 output. The API trade-offs: task-based async workflow with polling, USD-only billing, US hosting, and a narrow catalog (Runway models only). Railwail mirrors Gen-3 Alpha Turbo for compatibility and adds Veo-3 and WAN-2.5 as Google/Alibaba alternatives, all with synchronous calls and EU residency.

Step 1 — Get a Railwail API Key

Sign up at railwail.com and generate a key.

Step 2 — Replace Task-Based Polling

TypeScript / JavaScript

Before (Runway):

import { RunwayML } from "@runwayml/sdk";

const runway = new RunwayML({ apiKey: process.env.RUNWAYML_API_SECRET });

const task = await runway.imageToVideo.create({
  model: "gen3a_turbo",
  promptImage: "https://example.com/img.jpg",
  promptText: "a slow zoom out",
  duration: 5,
});

// Poll for completion
let result;
while (true) {
  result = await runway.tasks.retrieve(task.id);
  if (result.status === "SUCCEEDED") break;
  if (result.status === "FAILED") throw new Error(result.failure);
  await new Promise(r => setTimeout(r, 5000));
}
console.log(result.output[0]);

const res = await fetch("https://api.railwail.com/v1/videos", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.RAILWAIL_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "gen-3-alpha-turbo",
    prompt: "a slow zoom out",
    image_url: "https://example.com/img.jpg",
    duration_seconds: 5,
  }),
});
const json = await res.json();
console.log(json.video_url);

Python

import os, requests

resp = requests.post(
    "https://api.railwail.com/v1/videos",
    headers={"Authorization": f"Bearer {os.environ['RAILWAIL_API_KEY']}"},
    json={
        "model": "gen-3-alpha-turbo",
        "prompt": "a slow zoom out",
        "image_url": "https://example.com/img.jpg",
        "duration_seconds": 5,
    },
    timeout=300,
)
print(resp.json()["video_url"])

cURL

curl https://api.railwail.com/v1/videos \
  -H "Authorization: Bearer $RAILWAIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gen-3-alpha-turbo",
    "prompt": "a slow zoom out",
    "image_url": "https://example.com/img.jpg",
    "duration_seconds": 5
  }'

API Endpoint Mapping

Model Mapping

Pricing Comparison (May 2026)

Why Railwail Over Runway ML Direct

• EU hosting and EUR billing
• Synchronous responses — no task polling
• Multiple video models to A/B (Runway Gen-3, Veo-3, WAN-2.5)
• Same per-second pricing on Runway models
• Adds the rest of the AI stack (text, image, audio) to one key
• Per-key spend caps to prevent runaway video bills

FAQ

How long is the maximum video duration?

gen-3-alpha-turbo supports 5s or 10s. veo-3 supports up to 8s in the current API tier. For longer clips, generate multiple segments and concatenate.

Does the Runway Camera Motion control work?

Yes. Pass camera_motion: { type: 'pan-left', strength: 0.5 } as an extension parameter for Gen-3 models.

Can I do video-to-video?

Yes. Pass init_video_url to use an existing clip as the starting reference.

What about Runway's Video Upscaler?

Available as upscale-v1 via POST /v1/videos with mode: 'upscale' and source_video_url.

Webhook on completion supported?

Yes. Pass webhook_url in the request body. Railwail POSTs the completed response to that URL when ready.

Next Steps

• Sign up at railwail.com
• Generate an API key
• Replace Runway SDK with synchronous fetch / requests to /v1/videos
• Optionally try veo-3 or wan-2.5-i2v as alternatives
• Read the reference at railwail.com/docs
• Compare pricing at railwail.com/pricing