Decentralized
AI Compute
One OpenAI-compatible API for every AI model, served by the decentralized 0G Compute Network. Verified runs, privacy on demand, and zero platform fees.
One OpenAI-compatible API for every AI model, served by the decentralized 0G Compute Network. Verified runs, privacy on demand, and zero platform fees.
Any tool that speaks the OpenAI API works with 0G Compute. Change the base URL and API key, nothing else.
curl https://router-api.0g.ai/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer sk-YOUR_API_KEY" \
-d '{
"model": "glm-5.3",
"messages": [{"role": "user", "content": "Hello!"}]
}'Connect a wallet at pc.0g.ai, deposit 0G tokens or pay in USD, and create a key.
Set the base URL to https://router-api.0g.ai/v1 in the OpenAI or Anthropic SDK, or in Claude Code.
Verified routes only to attested providers. Private runs inside TEE enclaves so nobody sees your data, not even 0G.
Cloud AI services are expensive, opaque, and centralized.
Developers and enterprises pay premium prices, without visibility into execution, performance guarantees, or long-term ownership. A few infrastructure providers control nearly all AI compute, creating vendor lock-in and systemic risk.
GPU capacity is priced with heavy cloud markups and limited availability during demand spikes.
No way to verify that models ran correctly or that results weren't tampered with.
A few providers control all compute, creating vendor lock-in and single points of failure.
0G Compute combines decentralized infrastructure with proof-based verification to deliver trustless execution at scale.
Run inference and fine-tuning globally, without capacity constraints.
Every job comes with cryptographic guarantees of who ran it, how, and what was produced.
Pay per task, not per instance. Providers compete on cost and reliability, not brand reputation.
Plug into the 0G modular stack or connect to other chains and protocols.
0G Private Computer connects builders to on-demand AI services, all served by independent providers on the 0G Compute network.
Run LLMs, vision models, speech-to-text, and more through decentralized providers.
Customize models using your own datasets, with proof-based execution and onchain settlement.
Full training pipelines to unlock end-to-end AI development onchain.
0G Compute is not a single API, it is a modular execution system designed to support high-throughput AI workloads while preserving transparency and trust.
Onchain registration, payments, and settlement with cryptographic verification of every transaction.
Permissionless marketplace where independent GPU operators compete on reliability and cost.
One OpenAI-compatible endpoint with unified billing, automatic failover, and an SDK for wallet-signed dApps.
Proof-based execution with cryptographic guarantees you can independently audit.
From request to verified result, 0G Compute ensures every step is transparent and cryptographically secured.
Users submit inference requests through the 0G network. Smart contracts handle payment and route jobs to optimal providers.
Multiple verification mechanisms ensure results are trustworthy, from hardware attestations to zero-knowledge proofs.
0G Compute doesn't ask you to trust providers - it proves they executed correctly.
Hardware enclaves provide tamper-proof execution environments with remote attestation.
OPML allows fast results with economic guarantees against fraud.
ZKML provides mathematical certainty without revealing model weights or inputs.
Provider stakes ensure aligned incentives and slashing for misbehavior.
Read live from the 0G Compute Router model catalog
Turn your idle GPUs into a revenue stream. Join the 0G Compute network and earn tokens for every inference you serve.
Become a ProviderGet an API key at pc.0g.ai, then point any OpenAI-compatible client at https://router-api.0g.ai/v1. The router picks a provider, handles billing, and fails over automatically. If you are building a browser dApp where the user's wallet must sign each request, use the Direct SDK (Advanced mode) instead.
Verified routes only to attested providers: TeeML enclaves or official model channels verified via TeeTLS, so you always know which model ran. Private goes further and runs sealed inside TEE GPU enclaves, so nobody can see your data, not even 0G. Set the mode per API key or override it per request with a header.
Register your GPU with the 0G Compute network, stake the required tokens, and run the provider software. You'll automatically receive inference requests and earn rewards for successful completions.
Over a hundred models across chat, vision, video generation, speech-to-text, text-to-image, and embeddings, including DeepSeek, GLM, MiniMax, Gemini, and Claude families. Browse the live catalog with pricing at pc.0g.ai/models.
Providers set their own per-token rates and compete for requests, so there is no single-vendor markup, no top-up fee, and no premium for verified or private inference. You can pay in 0G tokens or in USD through Stripe. On average, users save 90% compared to cloud providers.
Whether you need AI compute power or have GPUs to share, 0G Compute connects you to the decentralized AI economy.
Get an API key and run verifiable AI inference at 90% lower cost than cloud providers.
Get API Key