How it works
From a strip on your timeline to a clip with a receipt
Three pieces work together: the add-on in Blender, a helper on your computer, and a GPU host you do not have to own or run.
You choose the work in Blender
Pick a tool in the Platformless AI panel of the Video Sequencer and select the strips it needs. The add-on places a placeholder where the result will go and shows the exact price.
The helper on your computer pays
A small helper program beside Blender holds your wallet or uses your card credits. Before paying, it checks the host’s published list of model templates against the record on-chain, so you know exactly what will run.
An independent GPU host renders it
Hosts are independent operators running open models: Lightricks’ LTX 2.3 and LTX 2.5, and NVIDIA Cosmos for VFX Passes. Your inputs are bound to the job, and the host renders the pinned template, nothing else.
The clip comes back with a receipt
The result is placed on your timeline and backed up, encrypted, to your own storage account. The payment settles on-chain against the work, and the clip keeps a record of the model, the settings and the payment.
Non-custodial account
Your account is a passkey on your own device. No server of ours ever holds or can rebuild your keys, and recovery is yours through your passkey, not a password reset.
No GPU needed
The rendering happens on the host, so any computer that runs Blender well can use every tool.
Open models
Every tool runs a published, pinned model template, so the same job always means the same thing.
Coming soon
Leave your email and we will send one message when it is ready to install.