AI model

Flux

Flux Schnell/Klein self-hosted image leaves give fast stills on Wavemaker GPUs with transparent node packs.

More than a Flux wrapper

Anyone can proxy an API. What you get here is what surrounds it:

  • Direction, not just generation — bind Flux into directed multi-scene spines (script → scenes → VO → music → assemble), or run the raw leaf when a single clip is all you need.
  • QC that can say no — directed graphs gate every clip and the final cut; failures re-roll or fail honestly instead of silently shipping.
  • Honest metering — GPU + partner meters shown before Confirm, unused reserve refunded, credits never expire.
  • Ownership — the graph is yours: edit it, publish it, earn per run.

FAQ

Partner API or GPU?

Flux leaves here are self-hosted GPU meters — see Nodes/Trust for pack license notes.

LoRA training?

Training Studio can train Flux-compatible LoRAs when Studio is enabled — separate surface.

Try link?

Try opens model-flux-schnell-t2i in the platform shell.