The decentralized
AI compute network
Gnodi turns thousands of operator-owned nodes into a distributed AI inference network. License a node and earn, build on an OpenAI-compatible API, or deploy enterprise-grade compute — settled on-chain in GNOD.
Built for Utility · Powered by Community
- 5,000+
- Network nodes
- 70B
- Max model params
- <1s
- Target P50 latency
- OpenAI
- Compatible API
Three ways in
One network. Three ways to plug in.
Gnodi aligns the people who supply compute, the developers who consume it, and the enterprises that depend on it — around a single decentralized inference network.
For Operators
License an AI node, serve inference from your hardware, and earn GNOD on every request — base rewards, per-inference fees, and uptime bonuses.
- 4 license tiers
- Earn per inference
- Founder NFT for early adopters
For Developers
Call open models through a REST + WebSocket API with an OpenAI-compatible endpoint. Python, JavaScript, and Go SDKs, plus LangChain and LlamaIndex drop-ins.
- OpenAI-compatible
- Streaming results
- Pay-per-inference in GNOD
For Enterprise
Reserve dedicated capacity with SLAs, custom fine-tuned models, and a dedicated relationship manager. Volume licensing and revenue-share options.
- 99.9% uptime SLA
- Custom model deployments
- Volume & custom terms
For Operators
License a node. Earn on every inference.
Licenses are priced in USD for predictable cost; rewards are paid in GNOD for network upside. Pay once, renew annually, and start serving inference in hours.
Light
$150 / yr renewal
Standard
$250 / yr renewal
Premium
$400 / yr renewal
Institutional
$650 / yr renewal
Existing Gnodi node operators get a 15% early-adopter discount plus a limited Founder NFT. Earnings shown are targets in GNOD and scale with token price and network demand.
For Developers
Decentralized inference, behind a familiar API.
Submit a prompt, the network routes it to a reputable node that supports your model, and you stream the result back. No single point of control, no per-seat lock-in.
Streaming inference
REST to submit, WebSocket to stream results token-by-token as nodes complete the work.
OpenAI-compatible
Point any tool at /v1/chat/completions for a drop-in replacement — no rewrites.
SDKs & frameworks
First-class Python, JavaScript, and Go clients, with LangChain and LlamaIndex integrations.
Settled on-chain
Pay per inference in GNOD. Payment escrows on submission; surplus refunds on completion.
import { GnodiClient } from "gnodi-sdk";
const gnodi = new GnodiClient(userAddress, signer);
// Submit an inference request — priced per call in GNOD
const task = await gnodi.infer(
"llama2-13b",
"Explain decentralized inference in one line.",
1.0 // max price in GNOD
);
// Stream the result as the assigned node produces it
const { result, actualPrice } = await gnodi.streamResults(task.taskId);
console.log(result, `— paid ${actualPrice} GNOD`);How it works
A three-layer network, end to end.
Every inference flows through the same path: a request is escrowed on-chain, routed to a qualified node, executed locally, then verified and paid — automatically.
Layer 1 · On-chain
Smart contracts
License NFTs, node registry & reputation, an escrowed task queue, and slashing — the trust layer that routes work and settles payment.
Layer 2 · Operator
Node daemon
A cross-platform Go daemon that wraps an inference engine, listens for tasks, runs the model, signs the result, and submits proof on-chain.
Layer 3 · Consumer
User-facing layer
A REST + WebSocket API, an operator dashboard, and SDKs that give developers a clean surface over the decentralized network underneath.
For Enterprise
Enterprise-grade AI compute, on your terms.
Reserve dedicated capacity with guaranteed performance, deploy custom models, and license at scale — backed by Block Reign and a dedicated relationship manager.
SLA-backed uptime
99.9% uptime guarantee, 1-hour priority support, and quarterly business reviews.
Custom & fine-tuned models
Deploy private or fine-tuned models to dedicated capacity across the network.
Volume licensing
Tiered discounts from 10 to 500+ licenses, or a revenue-share model for large operators.
The network
Economics built to last.
Gnodi is structured like a real infrastructure business: USD-priced licensing funds the network, GNOD rewards align everyone to its growth.
Real revenue, not inflation
Licenses are sold in USD and fund operations directly — no token emission to subsidize growth, so GNOD stays scarce as the network scales.
Aligned incentives
Operators pay predictable fiat and earn GNOD with upside. The treasury earns stable revenue. Demand for inference drives demand for the token.
Sustainable by design
Annual renewals and a DAO-funded budget for R&D, security, and growth — the model used by best-in-class infrastructure businesses.
Existing Gnodi community — founder licenses, first nodes live.
Crypto & GPU community — thousands of nodes, 10+ models.
Enterprise & cloud partnerships — 50K+ nodes, SLA tier.
Build the decentralized
AI compute layer with us.
Whether you run a node, ship an app, or scale an enterprise — there is a place for you on the Gnodi network.