Providing inference
A provider runs the Katara node on a machine with a supported GPU, serves one or more approved bundles, and is paid in USDC for every reply it completes. There is no stake and no minimum commitment.
What you need
Section titled “What you need”- Apple Silicon today: an M-series Mac with at least 24 GB of unified memory for the 8B bundle. Larger bundles state their own minimum.
- NVIDIA on Linux: a GPU that meets the bundle’s memory and compute capability, with Docker and the NVIDIA container toolkit installed. The node runs vLLM in a container. The first CUDA bundle profile is being pinned.
- A wallet on Katara (created in the portal) to register with and be paid to.
- A public HTTPS address the router can reach. Machines behind carrier-grade NAT need a tunnel; see Networking.
How it works
Section titled “How it works”- The node detects your hardware and shows which approved bundles it can run.
- You pick a bundle. The node fetches the pinned weights and checks every hash.
- You register once in the portal: your payout address, the node’s signing key, and its address.
- The node publishes signed price offers. The router sends you jobs when your offer wins.
- Each completed job returns a signed receipt. Katara settles it on chain and pays your address, less the 2% fee.
All five steps run today; the public test network admits providers as bundles are approved. Running a node has the commands.
Validation
Section titled “Validation”Validators send known prompts through the normal job path at intervals and score your answers. Providers that fail, stall, or return wrong results are held out of routing and, if it persists, deactivated. Serving exactly the pinned bundle keeps you eligible.
Your wallet key never touches the node. The node signs offers and receipts with a separate service key that you register alongside your wallet; rotate it from the portal at any time.
Read Running a node next.