Monetize your compute.
Contribute idle GPU power to the open network and earn rewards for serving verified inference requests.
Consumer Hardware Ready
Our verification engine is lightweight. You can run a node with high-end consumer GPUs or datacenter clusters.
Open Network
Open and permissionless by design. During testnet, operators are onboarded in curated waves as we scale capacity.
Automated Routing
Requests are automatically routed to your node based on availability and performance. No manual setup required.
Start serving in minutes
Minimum Hardware Requirements
Baseline specifications for node operators to ensure reliable network performance.
Node Setup
Initialize and start serving via our cross-platform CLI in minutes.
How it works
Provision Hardware
Set up your rig with at least one 24GB VRAM GPU and a stable connection.
Install Client
Run our cross-platform CLI. It handles weights, discovery, and networking.
Serve Inference
Receive requests, run the model, generate a proof of logits, and return it.
Earn Rewards
Get compensated directly for every successful verified request.
Frequently Asked Questions
How much can I earn?
Earnings are directly proportional to the number of verified inferences you serve. Your node's uptime, bandwidth, and the specific models you choose to serve will impact your total yield.
What tokens are used for payouts?
Providers are compensated in AMB tokens. Payments are distributed automatically to the wallet address registered during your node initialization.
Can I use multiple GPUs?
Yes. The Ambient CLI supports multi-GPU configurations for running larger models (like 70B+ parameters) or serving multiple concurrent requests across different GPUs.
Does running a node restrict my hardware?
No. You can start and stop the Ambient node process at any time. When you are not running the node, your hardware is entirely free for your own workloads.