Replicate

You send a request to a model through Replicate's cloud API, and you get back an image, a video clip, text or audio, depending on which model you picked.

Replicate is an AI API platform for running, fine-tuning and deploying open-source machine learning models without managing your own servers.

The catalog covers text-to-image models such as FLUX variants, Ideogram v3 and Recraft V3, video models like Google Veo 2 and Wan 2.1 image-to-video, and language models including DeepSeek R1 and Claude 3.7 Sonnet.

Billing is per second of compute, so the cost depends on the hardware your model runs on.

A CPU runs at $0.0001 per second, a T4 GPU at $0.000225, an A100 at $0.0014 and an H100 at $0.001525, with multi-GPU options up to 8x.

That works for bursty workloads, though long-running jobs on big GPUs add up quickly.

Pricing is freemium, so you can try models before committing.