Machine-payable inference for agents
Give an autonomous agent an OpenAI-compatible endpoint it can pay for per call in SOL — point the existing SDK at the gateway, keep your app code unchanged.
The agent economy makes far more inference calls than humans, at thinner margins, and increasingly with agents that hold their own wallets. Centralized APIs bill monthly against seat licenses and lock you to one base URL, key format, and pricing model — a poor fit for machine-to-machine payment, and expensive to unwind.
- The gateway exposes /v1/chat/completions and /v1/completions matching the OpenAI schema, so existing apps and agent frameworks switch with a base-URL change.
- Pass a Solana wallet address as the api_key — the wallet is both identity and payment, so an agent funds a session and pays per call with no human in the loop.
- Under the hood, requests can hit local PyTorch/Transformers, llama.cpp (GGUF), or proxied OpenAI, Anthropic, Cohere, or Ollama backends — one interface, swappable backends.
- Every call leaves an on-chain audit trail, so an agent (or its operator) can reconcile exactly what it paid for.
Try it for machine-payable inference for agents.
Read the deep dive, then run the quickstart.
Other use cases
Privacy-preserving inference
Route confidential prompts through onion circuits so no single node sees both source and content.
DePIN GPU compute that earns
Put idle DePIN or spot GPUs to work as an OpenAI-compatible inference node — dual SOL revenue from inference requests and federated training rounds.
Federated model training
Coordinate training across nodes that never share raw data, with FedAvg, FedProx, FedAdam, and SCAFFOLD.