fal
Image, video, and other generative-media APIs alongside managed deployment for custom workloads.
- Media-model APIs with model-specific parameters
- Managed serverless deployment for custom workloads
- Rates vary by model, output, and deployment type
fal provides hosted generative-media models and infrastructure for deploying custom inference workloads. Its catalogue includes image and video tasks with model-specific request parameters.
API or custom deployment
Hosted model endpoints avoid maintaining the underlying model servers. Custom serverless deployments offer control over an application workload, with scaling and execution monitoring. The output format, resolution, duration, and available controls depend on the chosen model.
How pricing works
Hosted endpoints use the model's listed unit, such as images, output dimensions, or generated duration. Custom deployments use their own compute arrangement. Estimate the cost using the intended resolution, clip length, and number of outputs. For another media API, see Runway .