GRIDLOCKAI that pays you backif it's too slow.
Pick a speed tier. If the AI responds too slowly, credits are automatically added to your balance—no waiting, no dispute, no support ticket.
Our technology stack
Read moreWhat is Gridlock, really?
Uber for AI compute, with refunds.
No one
controls it.
RunsonRobinhoodChain,anEVML2.Openworkerregistry,transparentjobreceipts,automaticSLArebates.
Inspired by ride-sharing efficiency. Built on trustless infrastructure. Priced by verifiable performance, not marketing.
A decentralized marketplace for AI compute , where accountability is automatic, not a support ticket.
Every request, receipted.
Job receipts, latency metrics, and SLA outcomes are logged in the explorer—transparent, auditable, verifiable.
Router-backed SLAs.
Miss the target and penalty credits land in your balance automatically—deducted from worker earnings, no dispute.
Incentives, not promises.
Workers are paid on goodput. Consumers are refunded on slack. The market pays for what it actually delivers.
The guarantee, in numbers.
How a request flows
Every call routes through the same five stages, from prompt to automatic settlement.
You send a prompt
Initiated via API or Console.
Router picks a worker
Selects the best GPU for the tier.
Worker runs the AI
The task is processed on the worker's GPU.
Response + proof
Result and verification delivered.
Penalty or payment
SLA misses credit your balance; hits pay the worker.
Powering decentralized AI , end to end.
A modern stack tuned for scale, security and verifiable performance. Every layer is instrumented, observable and swappable.
Every layer, accountable by design.
Gridlock
Robinhood Chain runs the receipts.
EVM L2 for smart contracts, tokenomics and automated SLA enforcement. Native ETH compatibility, every penalty and payout settled by a public contract.
One URL change. Guaranteed latency.
A drop-in replacement for the OpenAI API. Miss the SLA and you get automatic penalty credits, no tickets, no disputes.
- 1Point your SDK at api.grid-lock.tech
- 2Add sla: 'realtime' to your request
- 3Get paid back if it's late
import OpenAI from "openai";
const client = new OpenAI({
apiKey: process.env.GRID_KEY,
baseURL: "https://api.grid-lock.tech/v1",
});
const res = await client.chat.completions.create({
model: "llama-3.1-70b",
messages: [{ role: "user", content: "Hello" }],
// @ts-expect-error gridlock extension
gridlock: { sla: "realtime" },
});Pick how fast you need it.
Miss the target and penalty credits land in your balance automatically.
Chatbots, voice agents, live copilots.
RAG apps, copilots, general APIs.
Summarization, embeddings, backfills.
Enterprise · TEE · private prompts.
Two sides. One economy.
Every prompt funds a worker. Every second saved earns them more. Every missed deadline pays you back.
For builders who can't afford to be late.
- Guaranteed performanceFor chatbots, RAG apps and voice agents that need predictable latency.
- Cost efficiencyCompetitive pricing plus automatic refunds for underperformance.
- Confidential workloadsTEE-backed serving for privacy-sensitive prompts and data.
- Open by defaultCensorship-resistant, transparent infrastructure, no black boxes.
For operators who want to get paid for uptime.
- Monetize idle GPUsTurn underused hardware into passive income the moment you're online.
- Fair, automated payRewards settle from job fees, tied directly to SLA adherence and goodput.
- Own your uptimeSet your tiers, your hours, your stake. The router matches you to demand.
- Shape the networkGovernance rights scale with reputation, not just capital.
The only one backed bya guarantee.
The road to a verifiable AI economy.
Three phases. One direction: more capacity, more transparency, more control in the hands of the people running the network.
Core network stabilization & expansion.
Ship the foundation the rest of the ecosystem plugs into.
- 01Enhanced worker onboarding, join and configure in minutes.
- 02Granular SLA tiers for every class of workload.
- 03Expanded model catalog including open-source alternatives.
Developer tooling & integration.
Turn Gridlock into the fastest API to build against.
- 01Comprehensive SDKs across languages and runtimes.
- 02Community-contributed models with built-in monetization.
- 03Deep analytics dashboards for consumers and providers.
Decentralized governance & scalability.
Hand the keys to the network, and take it multi-chain.
- 01Full on-chain governance for token holders.
- 02L2 scaling to support massive transaction volumes.
- 03Cross-chain interoperability across major ecosystems.
Change one URL. Add one field. Get a latency guarantee backed by real money.
PENALTIES ARE AUTOMATIC · NO TRUST REQUIRED