The most flexible way to create. Nodes for prompts, references, models and results — drag a wire from one to the next and the whole chain stays in front of you, ready to re-run with a different model.
Every model runs on its official provider. You pay per generation, and every call is billed at least 20% below the provider's public price for the same call — some models go much lower. No seat, no tier, no wasted credits.
List price is the provider's own public API price for the same call (fal and Replicate charge the same). Prices per generation at the shown settings, in US dollars — longer clips and other resolutions scale the same way. Full price list for every model →
Top up once, generate whenever you want. Nothing monthly, nothing expires, and every failed generation refunds itself.
freym also tracks what creators post on Threads and X, with the full prompt when they share it. Copy a prompt, paste it into the canvas, run it on any model.
freym reads the official accounts of every gen-AI lab and keeps the releases, not the noise. New models land in the canvas soon after they land here.