Hybrid compute model
The product narrative connects AI inference capacity planning, API routing, and B2B demand. It does not claim that a specific physical GPU belongs to an individual user.
Product architecture
Compute Nodes for AI revenue sharing
Understand the LLMsRelay Compute Node product model for inference capacity planning, B2B API traffic, ledger accruals, and 90-day allocations.
Direct answer
A Compute Node is the accounting unit used by LLMsRelay Earn. It records the user's allocation, rate, lifecycle, accrual ledger, and exit terms without presenting unverified hardware telemetry.
Node families
Micro, Pro, Enterprise, and Cluster families make the amount curve easier to understand while preserving every rate level from 10% to 40%.
Auditable lifecycle
Pending, active, liquidation requested, completed, and liquidated states make the operational workflow visible to users and administrators.
Frequently asked questions
Does a Compute Node represent a specific GPU?
No. It is an allocation and accounting unit for the Earn workflow, not a title to a specific graphics card.
Where are accruals recorded?
Each credited amount is stored as an Earn ledger entry associated with the user and node.
Are GPU provider integrations required?
No. The current backend operates independently of GPU-provider integrations.
Related Earn guides
Start with your amount
Review the projection and submit an activation request from your account.