A single endpoint for AI image and video generation. Send one prompt and a few parameters, pick as many models as you like, and get back every output at once, images and videos, generated in parallel across 105 models.
The Generation API lets you create images and videos from a single API endpoint. Instead of integrating dozens of providers and learning a different request format for each model, you talk to one endpoint with one consistent set of parameters. Choose one model or many, for every model you pick, you get an output. It's the fastest way to compare models, produce variations, or build an entire batch of creative from a single request.
Generate both from the same endpoint, text-to-image, text-to-video, and image-to-video.
List several models and the API runs them in parallel, returning one result per model.
A single command with a few parameters produces any number of outputs across models.
All results come back bundled, one ZIP of images and one ZIP of videos, or via webhook.
The key idea: with a single endpoint, one command and a few parameters, you can generate as many images and videos as you need, across different models, in one go.
Create an API key in your dashboard. Authenticate with a simple x-api-key header (or a self-contained URL).
Choose your video and image models and set shared parameters, prompt, aspect ratio, resolution, duration, quality. Values are auto-adjusted to what each model supports.
The API generates with all chosen models in parallel and returns the finished media as ZIP files, or calls your webhook the moment it's ready.
One integration replaces dozens. You stop maintaining a separate client, request format, and retry logic for every model, and gain the ability to fan out a single prompt across many models instantly.
Add a model by name, no new SDK, no new auth, no new code. New models become available without changing your integration.
Every chosen model runs at the same time under a managed concurrency limit, so a big batch finishes fast.
Send standard values once; we translate them to each model's own fields and snap anything unsupported to the nearest valid option.
Async jobs, webhooks, automatic batching/queueing, and tidy ZIP delivery, ready for production workloads.
Pick the right tool for each job. Below are the families we support and what they're best at, followed by the complete, live list of every available model.
Cinematic, true-to-life video with strong prompt adherence and natural audio, great for premium, story-driven clips.
Lifelike motion and physics with excellent image-to-video, ideal for animating a still into realistic movement.
Fast, flexible video with a wide range of durations (4–15s), perfect for quick iterations and social formats.
High-fidelity, coherent video for complex scenes and motion, for when realism and consistency matter most.
Versatile generators with reference-to-video options, strong for stylized looks and character-consistent clips.
Creative and stylized video engines, effects, anime, dreamlike motion and fast pro-grade output.
High-quality images with excellent text rendering and prompt control, great for ads, posters and UI mockups.
Native 4K, multi-subject consistency and extreme aspect ratios, ideal for crisp, detailed, brand-accurate visuals.
Artistic, photoreal and richly detailed image generation across a wide range of styles.
Every model below can be used by its name in a request. Start-frame models (image-to-video / image-to-image) are marked ▶ frame.
Generation runs on credits. Every image or video you create with a model consumes credits, heavier models (like premium video) cost more, lighter ones (like images) cost less. Buy a package and spend it across any image or video model you like.
Credits are consumed only when you generate. Different models use different amounts of credits per output, your dashboard tracks usage and cost per model.
Create an API key and build your first multi-model request in minutes.