Spindle
ASSEMBLING THE NETWORK001
COMPUTE / VERIFY / CHOOSE

About Spindle

We connect GPU work to a payout you can choose

Explore network roles

The idea

02

Compute lanes

03

Payout routes

M1

Specification build order

Build sequence

How the design fits together

M1

Node identity

The specification begins with bonded node identity, registry contracts and an open-lane agent.

M2

Receipt roots

Verified work enters an earnings ledger, with cumulative Merkle roots posted in hourly epochs.

M4

Payout routes

Claim USDG, then quote a Stock Token or fund a card with your wallet.

M6

Sealed compute

In testing: confidential GPUs add an attested sealed lane for encrypted prompts and responses.

Values

What guides the design, from node to claim

Show the receipt

Every job emits a signed work record. Receipts support metering; they do not prove output correctness.

Privacy has a defined lane

Confidential compute belongs in the sealed lane. Open-lane jobs must be non-sensitive.

Keep payout choice clear

Earnings accrue in USDG. Providers claim to a wallet, then pick a Stock Token or card.

Make limits visible

Buyer demand, supported regions, card terms and Stock Token liquidity shape what is available.

Core network roles

The roles behind each job

GPU providers

Supply eligible compute

AI builders

Bring suitable workloads

Node operators

Maintain node performance

Developers

Build with the API

NETWORK DESIGN

Inside Spindle

Network roles

Explore the network roles