Runway API Guide: Features, Pricing & Code (2026)

Runway makes some of the most controllable AI video on the market, and Gen-4.5 currently sits at the top of several third-party video benchmarks. This guide covers what the Runway API actually does, what it costs, how to get access, and a working code example to generate your first video.

What is the Runway API

The Runway API gives developers access through code to Runway's video generation models, currently Gen-4 Turbo and Gen-4.5, without going through the consumer web app. It supports both text-to-video and image-to-video generation: give it a prompt alone, or a prompt plus a starting image, and it renders a clip from there.

The API is video-focused. There's no dedicated image generation endpoint tied to these two models the way there is for, say, GPT Image or Nano Banana, and Runway's API doesn't offer full music generation the way Suno does (it does expose separate audio endpoints for text-to-speech and sound effects, just not song generation). If you need image generation outside these two video models, or music generation specifically, you'll be pairing Runway with a separate model or provider for those.

Runway API features

Text-to-video and image-to-video. Both Gen-4 Turbo and Gen-4.5 accept a text prompt on its own, or a prompt combined with a reference image that becomes the starting frame of the clip. Image-to-video is generally the more reliable path when you need a specific subject, character, or scene to carry through consistently, since the model has a concrete starting point rather than generating one from text alone.

Camera movement and style consistency. Runway's reputation for control comes largely from the web app, where features like Motion Brush let you paint specific regions of an image to direct movement point by point, and dedicated camera controls let you dial in pans, arcs, and zooms. Through the API, that control is mostly expressed through prompt language rather than as separate settings. Camera movement described in the prompt (a slow dolly-in, an arc around the subject) tends to carry through reliably, but the Motion Brush feature itself is a web-app feature, not available as an API option today. If your use case depends specifically on Motion Brush, check the current API parameter reference before assuming it's available.

Resolution, duration, and output limits. Both models generate at 720p and support 5-second or 10-second clips. Available aspect ratios include 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9, so you can target landscape, portrait, or square output without cropping after the fact. Neither model currently generates native audio through the API; if you need sound, you're pairing the video output with a separate audio or music generation step.

Runway API pricing

Official Runway credit system explained

Runway's own developer API lives at a separate portal, dev.runwayml.com, with its own credit pool that's completely separate from consumer Standard, Pro, or Max subscription credits. Credits cost $0.01 each, and you need a minimum $10 top-up before your first API call. There's no waitlist or approval step for the API itself, just sign-up, add credits, and go.

Cost per second by model (Gen-4 vs Gen-4.5)

On Runway's own API, Gen-4 Turbo runs 5 credits per second ($0.05/sec), so a 5-second clip costs $0.25 and a 10-second clip costs $0.50. Gen-4.5, the flagship model, runs 12 credits per second ($0.12/sec): $0.60 for 5 seconds, $1.20 for 10 seconds.

Where third-party providers differ on price and why

Multi-model API platforms that bundle Runway alongside other video, image, and music models typically charge more per second than Runway's own API, because you're paying for the convenience of one integration and one bill across many models rather than managing access with each provider separately. On Apiframe, for example, Gen-4 Turbo costs 36 credits for 5 seconds or 73 for 10 (about $0.36 and $0.73 at the pay-as-you-go rate), and Gen-4.5 costs 87 credits for 5 seconds or 174 for 10 (about $0.87 and $1.74). That's a real markup over going direct, but it buys you the same request and webhook format you'd use for Kling, Sora, Seedance, Veo, and dozens of other models, instead of learning a new schema for each one. Other providers land at different points on that same tradeoff: some flat-rate resellers charge less than Runway's own API for narrower feature sets (image-to-video only, for instance), while others charge more for broader compatibility. It's worth comparing a few before committing, since the "best" option depends on whether you need just Runway or a whole catalog of models behind one key.

How to get Runway API access

Official API waitlist/access process vs aggregator access

Runway's own API isn't gated behind a waitlist or sales call, which surprises people who expect it to work like some other closed video models. Setup is three steps: create an account at the developer portal, create an "organization" (which holds your keys and billing config), then add the minimum $10 in credits before your first request.

Going through a multi-model API platform like Apiframe skips the separate developer-portal account entirely. You sign up once, get an API key from the dashboard, and Runway's models sit alongside every other model on the same key and the same credit balance.

API key setup and authentication

On Runway's own portal, you create a key under the API Keys tab, copy it immediately (it's shown once), and pass it via the RUNWAYML_API_SECRET environment variable, which Runway's official SDKs pick up automatically.

On Apiframe, you generate a key from the dashboard; keys start with the prefix afk_, and you pass yours in the X-API-Key header on every request. New accounts start with 100 free credits, enough to test a single 5-second Gen-4 Turbo clip before you need to add funds.

Runway API code example

Making your first generation request (Python)

Here's a text-to-video request against Apiframe's unified endpoint, one of the two ways to reach Runway's models programmatically:

import requests

response = requests.post(
    "https://api.apiframe.ai/v2/videos/generate",
    headers={
        "X-API-Key": "afk_your_api_key_here",
        "Content-Type": "application/json",
    },
    json={
        "prompt": "a cinematic slow-motion shot of ocean waves crashing on volcanic rocks",
        "model": "runway-gen4.5",
        "runwayParams": {
            "duration": 10,
            "aspect_ratio": "16:9",
        },
    },
)
data = response.json()
print(data)  # {"jobId": "...", "status": "QUEUED"}

Swap "model": "runway-gen4.5" for "runway-gen4-turbo" if you want the faster, cheaper tier instead of the flagship. For image-to-video, add an image field inside runwayParams pointing to a publicly accessible image URL, and the model uses it as the starting frame instead of generating one from scratch.

Polling for the finished video and handling errors

Generation runs in the background: the request above returns a job ID right away, and you poll a separate endpoint to check on it.

import time
import requests

job_id = data["jobId"]
headers = {"X-API-Key": "afk_your_api_key_here"}

while True:
    job = requests.get(f"https://api.apiframe.ai/v2/jobs/{job_id}", headers=headers).json()
    status = job["status"]

    if status == "COMPLETED":
        print(job["result"]["videoUrl"])
        break
    elif status == "FAILED":
        print("Generation failed:", job.get("error"))
        break

    time.sleep(5)

Jobs move through QUEUED, then PROCESSING (with a progress percentage you can show to users), then either COMPLETED or FAILED. If a job fails, credits are refunded automatically, so a failed generation doesn't cost you anything. Watch for 402 (insufficient credits), 429 (rate limit exceeded), and 400 (a validation error, usually a malformed runwayParams value) in your error handling, alongside the usual 401 for a missing or invalid API key. If you'd rather not poll, both Runway's own API and most aggregators support webhooks that notify your server when a job finishes.

Runway API vs alternatives

Pricing below is per second at Apiframe's standard tier, which lets you compare Runway against other video models on the same billing basis. Figures for other providers' own direct APIs will differ.

Runway's strongest differentiator isn't raw price, it's control: prompt-driven camera movement and shot-to-shot style consistency that make it a common pick for film-adjacent and branded creative work, where getting a specific look matters more than generating the cheapest possible clip. If your priority is lowest cost per second or native 4K, Seedance or Kling's standard tier will usually beat Runway on a pure price basis.

For a broader comparison of where to access Runway, including other third-party providers beyond Apiframe, see our Best Runway API Providers Compared guide. You can also explore how Runway stacks up in our Best AI Video Generation APIs roundup, or see how prices compare across models in AI Video API Pricing in 2026.