AI video generation API
Text to video and image to video through one endpoint, with the newest releases first — Kling, Seedance, Wan, Grok Imagine and more. Submit a job, poll for completion, download the result. No GPU cluster, no queue to operate.
Most recent addition: Wan 3.0, added Sep 2026. We add the best video models available, each behind the request shape you already use.
Pay per request. Powering 1M+ generations daily.
import GetimgAI from "getimg-ai";
const client = new GetimgAI();
const submission = await client.videos.generations.create({
model: "wan-3-0",
prompt: "A realistic vertical commercial for a pair of black-and-white canvas skate shoes, filmed in a clean concrete skatepark at late afternoon. Open with a macro close-up of the rubber sole touching the ground, followed by detailed shots of the canvas texture, laces and side profile. A skater pushes forward, lands one clean trick and rides past the camera. Keep the product visible and recognizable in every shot. Use low tracking angles, short controlled handheld movement, natural sunlight, hard concrete shadows and restrained black, off-white and muted red accents. The campaign should feel youthful, raw and confident, but still polished and commercially usable. End with the shoes resting on the edge of the ramp in a simple hero shot.",
aspect_ratio: "16:9",
resolution: "480p",
duration: 4
});
const requestId = submission.id;
// Poll
let result;
while (true) {
result = await client.videos.generations.retrieve(requestId);
if (result.status === "completed") break;
if (result.status === "failed") {
throw new Error(result.error.message);
}
await new Promise((r) => setTimeout(r, 5000));
}
console.log(result.data[0].url);Every video model, compared
Live pricing, clip length, audio support and accepted inputs, pulled straight from the same catalog the API serves. Newest first.
| Model | Added | From / second | Max resolution | Max length | Inputs |
|---|---|---|---|---|---|
| Wan 3.0alibaba | Sep 2026 | $0.05/second | 1080p | 30s | First frame, Last frame, Reference |
| MiniMax H3 Maxminimax | Sep 2026 | $0.06/second | 720p | 15s | First frame, Last frame |
| Grok Imagine Video 1.5xai | Aug 2026 | $0.09/second | 1080p | 15s | First frame |
| Gemini Omni Flash 1.1google | Aug 2026 | $0.033/second | 4K | 10s | First frame, Last frame, Reference |
| MiniMax H3minimax | Aug 2026 | $0.15/second | 2K | 15s | First frame, Last frame, Reference |
| Gemini Omni Flashgoogle | Jun 2026 | $0.11/second | 720p | 10s | First frame, Reference |
| HappyHorse 1.1alibaba | Jun 2026 | $0.14/second | 1080p | 15s | First frame, Reference |
| Kling 3.0 Turbokling | Jun 2026 | $0.112/second | 1080p | 15s | First frame |
| Grok Imagine Videoxai | Jun 2026 | $0.06/second | 720p | 15s | First frame, Reference |
| HappyHorse 1alibaba | Apr 2026 | $0.14/second | 1080p | 15s | First frame, Reference |
| Wan 2.7alibaba | Apr 2026 | $0.10/second | 1080p | 15s | First frame, Last frame, Reference |
| Kling O3kling | Feb 2026 | $0.084/second | 4K | 15s | First frame, Last frame, Reference |
| Kling 3.0 Prokling | Feb 2026 | $0.084/second | 4K | 15s | First frame, Last frame |
| Seedance 1.5 Probytedance | Dec 2025 | $0.026/second | 1080p | 12s | First frame, Last frame |
| Wan 2.6alibaba | Dec 2025 | $0.10/second | 1080p | 10s | First frame |
| Seedance 1.0 Pro Fastbytedance | Nov 2025 | $0.022/second | 1080p | 12s | First frame |
Which video model should you call?
Generation quality in video has moved faster than in images, so the release date matters more here than anywhere else in the catalog.
Start here — newest model
Wan 3.0, added Sep 2026. Motion coherence, physics and prompt adherence improve sharply between generations, so the newest release is usually the safest default.
Image to video
Wan 3.0 accepts both a first and last frame, so you can animate an existing asset and control where the shot lands instead of hoping the model guesses.
Longest single clip
Wan 3.0 runs up to 30 seconds in one request, which avoids stitching and the continuity errors that come with it.
Start in 60 seconds
Install the official SDK for Node.js or Python, drop in your API key, and submit your first video job. Video generation is asynchronous — you poll the status endpoint and download the result when it completes.
npm install getimg-ai
import GetimgAI from "getimg-ai";
const client = new GetimgAI();
const submission = await client.videos.generations.create({
model: "wan-3-0",
prompt: "A realistic vertical commercial for a pair of black-and-white canvas skate shoes, filmed in a clean concrete skatepark at late afternoon. Open with a macro close-up of the rubber sole touching the ground, followed by detailed shots of the canvas texture, laces and side profile. A skater pushes forward, lands one clean trick and rides past the camera. Keep the product visible and recognizable in every shot. Use low tracking angles, short controlled handheld movement, natural sunlight, hard concrete shadows and restrained black, off-white and muted red accents. The campaign should feel youthful, raw and confident, but still polished and commercially usable. End with the shoes resting on the edge of the ramp in a simple hero shot.",
aspect_ratio: "16:9",
resolution: "480p",
duration: 4
});
const requestId = submission.id;
// Poll
let result;
while (true) {
result = await client.videos.generations.retrieve(requestId);
if (result.status === "completed") break;
if (result.status === "failed") {
throw new Error(result.error.message);
}
await new Promise((r) => setTimeout(r, 5000));
}
console.log(result.data[0].url);Video generation API FAQ
What is an AI video API?
It is an HTTP endpoint that turns a prompt, or a still image, into a video clip. You POST a model id, a prompt, a duration and a resolution, then poll for the finished file.
Unlike image generation, video takes long enough that the request cannot stay open, so the API is asynchronous by design.
Which video models are available through the API?
16 video models are live right now. The most recent additions are Wan 3.0, MiniMax H3 Max, Grok Imagine Video 1.5. Each has its own page with parameters, code samples and pricing.
Generating stills instead? See the AI image generator API.
How do you decide which video models to add?
We add the best video models available. Each candidate is assessed on motion quality, reliability and cost before it goes into the catalog.
Once a model is in, it is callable immediately — same endpoint, same request format, no SDK upgrade. Only the model id changes.
The most recent addition is Wan 3.0, added Sep 2026.
Can the API generate video with sound?
Some models generate synced audio alongside the picture, others are silent. Support is per model — check the model page for the one you plan to call, which lists its exact parameters.
How much does a video generation API cost?
Video is billed per second of output. Current video rates run from $0.022 to $0.42 per second, so a five second clip lands between $0.11 and $2.10 depending on model, resolution and audio. There is no subscription and no minimum spend.
What is the difference between text to video and image to video?
Text to video generates a clip from a written prompt alone. Image to video animates a still you supply as the first frame, which gives you far more control over composition, character and brand consistency.
Both run through the same endpoint — image to video just adds a reference frame to the request. Several models also accept a last frame so you can pin where the shot ends.