# 0G Builder Hub (full) 0G is the trust layer for AI. It combines private compute, encrypted storage, and an AI-native chain in one stack where AI work, from inference to agent actions, can be verified onchain. This file is the expanded, machine-readable digest of https://build.0g.ai. Each ## section is a page or content set with its canonical URL under the heading. The concise index lives at https://build.0g.ai/llms.txt. The Ask AI assistant lives at https://build.0g.ai/ask and runs on 0G Compute. --- ## Home https://build.0g.ai Top-level entry point. Hero: "Build the next generation of onchain AI on the trust layer for AI." The stack has four surfaces: Compute, Storage, Chain, and Agentic ID, presented as components of one stack. Sections: Stack, Resources (Tools, Get Funded, SDKs, Documentation, Tutorials, Showcase), Builds. --- ## Native token naming (CANONICAL) The native gas token across both 0G testnet and 0G mainnet is 0G (no longer A0GI). Earlier docs, READMEs, and npm metadata still reference the old name A0GI (and wA0GI for the wrapped variant); those references are out of date. Always say "0G tokens" or "0G (native gas token)", never "A0GI" or "AOGI"; "Wrapped 0G" or "w0G", never "wA0GI". --- ## Network config (CANONICAL VALUES) https://build.0g.ai/chain Galileo Testnet - Chain name: 0G-Galileo-Testnet - Chain ID: 16602 (hex 0x40EA) - RPC: https://evmrpc-testnet.0g.ai - Explorer: https://chainscan-galileo.0g.ai - Faucet: https://faucet.0g.ai - Native token: 0G, 18 decimals Mainnet - Chain name: 0G Mainnet - Chain ID: 16661 (hex 0x412D) - RPC: https://evmrpc.0g.ai - Explorer: https://chainscan.0g.ai - Native token: 0G, 18 decimals Deployment: any Solidity contract compiled for the cancun hard fork with Solidity 0.8.19 deploys unchanged. Pin evmVersion: cancun; newer EVM versions may not verify on the explorer. --- ## Get Funded https://build.0g.ai/funding Three different balances, constantly conflated: 1. Testnet gas: free 0G on Galileo. Primary faucet https://faucet.0g.ai; fallback Google Cloud faucet https://cloud.google.com/application/web3/faucet/0g/galileo when rate-limited. Testnet 0G has no market value. 2. Mainnet 0G: real tokens. Canonical route is https://get.0g.ai, an interactive decision tree that routes by starting point (fiat, crypto on another chain, CEX funds, DeFi, existing wallet). Exchange and bridge listings live there, not here. 3. Compute credit: the spendable balance that pays providers for inference on 0G Compute. Credit is not gas. Top up at https://pc.0g.ai, or via CLI: `0g-compute-cli deposit --amount 3` (3 OG minimum, creates the ledger). Check balance with `0g-compute-cli get-account` or the PC dashboard. --- ## Compute https://build.0g.ai/compute Decentralized AI inference over a GPU marketplace. Open provider bidding sets the price, you pay per request, settlement happens on chain. ### Primary path: Private Computer Private Computer is the fastest way onto 0G Compute. One API key gets you the whole model catalog: chat, image, speech, and embeddings, through an OpenAI-compatible endpoint, so you change a base URL and keep your existing SDK. Open provider bidding sets the price, and there are no platform or top-up fees. Every request is verifiably routed, and you can dial privacy up per request when the workload needs it. - Get an API key: https://pc.0g.ai - Model catalog with pricing and trust-mode support: https://pc.0g.ai/models - Browser playground: https://pc.0g.ai/playground - Live request/token/TEE-proof counters: https://pc.0g.ai/stats ### Advanced path: wallet auth with the broker (CLI quickstart) 1. Install: `npm add @0gfoundation/0g-compute-ts-sdk -g` and `npm install openai` (Node 20+) 2. Provision: `0g-compute-cli setup-network`, `0g-compute-cli login`, `0g-compute-cli deposit --amount 3` (3 OG minimum), `0g-compute-cli inference list-providers` 3. Copy a provider address from the list-providers output: `export PROVIDER=`, then `transfer-fund`, `inference acknowledge-provider`, `inference get-secret` (prints an app-sk-... secret and service URL) 4. Call with the OpenAI SDK: baseURL `${ZG_SERVICE_URL}/v1/proxy`, apiKey = the app-sk secret, model `0GM-1.0-35B-A3B` (or any model listed on pc.0g.ai/models) 5. Balance: `0g-compute-cli get-account` or pc.0g.ai The CLI wraps the broker SDK (`createZGComputeNetworkBroker` from @0gfoundation/0g-compute-ts-sdk) for in-app wallet-auth flows. ### Trust modes Not every task needs the same guarantee. Standard is the default and has the widest coverage; step up only when the workload calls for it. | Mode | What it gives you | Use when | |---|---|---| | Standard (default) | Widest coverage: every model and provider including community channels. Routing is not attested. | Everyday tasks, best price and availability | | Verified | Routing restricted to attested providers. You get a signed proof that the model you called is the model that ran, no silent substitution. | You need provenance | | Private | TEE-isolated execution path. Neither 0G nor the provider sees the prompt. | Sensitive data | How to set it: - Per request: header `X-0G-Provider-Trust-Mode: verified|private` - Per key default: PC Dashboard, API Key, then Edit ### Attestation primitives | Primitive | What it proves | Qualifies for | |---|---|---| | TeeML | The model runs inside the enclave. Hardware baseline is an Intel TDX CPU plus an NVIDIA H100/H200 with TEE support. The prompt enters the enclave encrypted, the output is signed inside the enclave before it leaves, and the host machine only ever sees encrypted traffic. | Verified and Private | | TeeTLS | The routing path, not the execution. A Broker running inside a TEE proxies the request to the provider over HTTPS, pins the provider's TLS certificate during the handshake, then signs a routing proof bundling the cert fingerprint, request hash, response hash and provider identity. Conceptually close to zkTLS with stronger privacy, since the TEE guarantees the relay itself is trustworthy. | Verified only: inference still happens at the upstream provider, which can see the prompt | | SPML | For open-source models that are neither self-hosted in a 0G enclave nor served from a closed-source official channel: verifies the provider's GPU and model deployment. | Verified | Accuracy notes (important): - Most models are NOT running inside a TEE. Only a minority of providers are TeeML; most are TeeTLS, which attests the routing path and not the execution. Never claim "your model runs inside a TEE" as a blanket statement. - Standard is not the absence of a guarantee. If a model is served only by TeeML providers, requests run inside an enclave whichever mode is picked. --- ## Storage https://build.0g.ai/storage Fast, decentralized storage for AI workloads. A Log layer for append-only archives, a Key-Value layer for live state, 30+ MB/s retrieval measured on mainnet, at $11/TB/month on Turbo. Quickstart: install `@0gfoundation/0g-storage-ts-sdk`, get testnet tokens (faucet.0g.ai; see /funding), initialize signer + indexer, upload, retrieve, verify on Storage Scan. Starter kits: Storage Web (browser + wallet), Node TypeScript, and Go variants. Explorers: https://storagescan.0g.ai (mainnet), https://storagescan-galileo.0g.ai (testnet). --- ## Chain https://build.0g.ai/chain EVM-compatible L1 tuned for AI apps, built on CometBFT with sub-second finality. Deploy with Hardhat or Foundry using the network config above. Ready-made deployment scripts: the 0G deployment scripts starter kit. Index and stream chain data with Goldsky: https://docs.0g.ai/developer-hub/building-on-0g/indexing/goldsky --- ## Agentic ID https://build.0g.ai/agentic-id Onchain identity for AI agents, built on ERC-7857. Mint and manage agent identities; examples in the agenticID-examples repo. --- ## SDKs & Starter Kits https://build.0g.ai/sdks Official SDKs and starter kits for Storage, Compute, and Agentic ID: - 0G TypeScript SDK (`@0gfoundation/0g-ts-sdk`) - Compute broker SDK and CLI (`@0gfoundation/0g-compute-ts-sdk`) - Storage Rust SDK/CLI (`0g-storage-rust-sdk`) - Storage TypeScript SDK (`@0gfoundation/0g-storage-ts-sdk`) and Go starter kit - Compute TypeScript starter kit, Storage Web starter kit, 0G deployment scripts --- ## Tools https://build.0g.ai/tools Essential: Testnet Faucet (faucet.0g.ai), Storage Scan Mainnet (storagescan.0g.ai), Storage Scan Testnet (storagescan-galileo.0g.ai), Chain Scan Mainnet (chainscan.0g.ai), Chain Scan Testnet (chainscan-galileo.0g.ai), Private Computer (pc.0g.ai), 0G Hub (hub.0g.ai: liquidity and bridging front end with multi-hop swap routing and live-quoted end-to-end bridging). Apps: 0G App (app.0g.ai), Genome (app.0g.ai/genome), 0G Chat (app.0g.ai/private-chat), Agent Launchpad (app.0g.ai/agent-launchpad), PC Playground (pc.0g.ai/playground). Infrastructure: QuickNode RPC, Ankr RPC. Community: OpenAdapter, 0G Agent Skills (github.com/0gfoundation/0g-agent-skills), 0G Code to Coin MCP server (npmjs.com/package/@0gfoundation/0g-cc), 0g-mcp-server (github.com/colygon/0g-mcp-server), 0g-chain-mcp (github.com/Tairon-ai/0g-chain-mcp), Goldsky indexing. --- ## Tutorials https://build.0g.ai/tutorials Video walkthroughs and office hours. Recent highlights: - ETHGlobal Lisbon Workshop: 0G, Permissionless Blockchain for AI (16:29): https://youtube.com/watch?v=HKNix-5qD9s - Build Agents with 0G + Office Hours (18:37): https://youtube.com/watch?v=ViyYRWgzdVs - ETHGlobal Cannes Workshop: Building with 0G Storage and Compute (14:17): https://youtube.com/watch?v=LxcV-DGmNLs - 0G & Goldsky: Indexing and Streaming 0G Data: https://youtube.com/watch?v=nCn3atlqkpk Plus dev guides for each product and recurring office hours. Full list on the page. --- ## Build https://build.0g.ai/build The starting point for builders: `curl -fsSL https://build.0g.ai/install | sh` sets up the 0G agent skills for your coding assistant, then the page indexes the four products (Compute, Storage, Chain, Agentic ID) and the resources around them (Tools, Get Funded, SDKs, Documentation, Tutorials, Showcase). --- ## Zero Coding https://build.0g.ai/agents Build on 0G without writing code: point an AI coding assistant at 0G with prompt templates per product (Storage, Chain, Compute, Agentic ID, Full Stack) plus resources including AI Context Documentation, 0G Compute skills for Claude Code, 0G Agent Skills, and the 0g-cc MCP server (`claude mcp add 0g-cc npx @0gfoundation/0g-cc`). Agents: this llms-full.txt (and /llms.txt) is the machine-readable entry point to the hub. --- ## Documentation index https://build.0g.ai/documentation Sections: Getting Started, Concepts, 0G Compute, 0G Storage, Agentic ID. Curated links into https://docs.0g.ai. --- ## Builders Update https://build.0g.ai/updates The builder changelog. Every week or two an update "block" is sealed: network/protocol news, new features across the stack, events, builder spotlights, and curated AI news. Append-only, deep-linkable per block (#block-N). /blocks redirects here. --- ## 0G Dev Day 2026 https://build.0g.ai/dev-day-2026 0G Dev Day 2026, "Zero to Infinity": Oct 5, 2026 at National Gallery Singapore (main summit 10:00-17:00 SGT, VIP reception 19:00-22:00 at Arts House). Four conference tracks (Trustless AI Agents; AI Trust Layer; DeFi / Privacy / Financial Systems; Ecosystem Builders), 20+ speakers, demos, build sessions and a hackathon. 2700+ registrations, 10 ecosystem booths and 7 experiences (massage stations, F1 simulator, barber shops, nail salon, photo booths). Register at https://luma.com/0g-devday-2026. Four ways to earn 0G AI Credits on the day: - 0G Memo Mystery Box (https://build.0g.ai/dev-day-2026/memo): collect a 0G Momo Mystery Box at check-in (a Labubu character and a companion card), photograph your Labubu, scan the card's QR code to claim 0G PC Credits at pc.0g.ai, then post on X and show the post for an official 0G stamp. The 0G Momo guardians: 0GMemo (memory), 0GNome (creativity), 0GPrico (privacy). - Train a Robot Dog (https://build.0g.ai/dev-day-2026/robot-dog): train a robot dog, record the session and upload it to X; AI Credits for a verified submission. - Hacker Zone (https://build.0g.ai/dev-day-2026/hacker-zone): the build stations' launchpad. Every machine has Claude Code and OpenCode running on 0G models (0GM-1.0-35B-A3B, GLM-5.2), the 0G agent skills in ./.0g-skills, a 0G Compute API key (OpenAI-compatible endpoint https://router-api.0g.ai/v1) and a funded Galileo testnet wallet. Pick a build idea (~10 min Compute-only, ~20 min two-piece, or stretch), copy its prompt, paste it into the terminal, ship to GitHub; AI Credits for valid participation. - Talks: attend the conference tracks. --- ## Testing Program https://build.0g.ai/testing Sign up with a wallet, test 0G apps and infrastructure, report reproducible bugs, earn up to 100 0G in Compute Credit across levels L0-L3. --- ## Showcase https://build.0g.ai/japan is the Japanese community page; https://build.0g.ai/showcase lists builder projects across hackathons, buildathons, and Zero Coding events. --- ## Ask AI https://build.0g.ai/ask Chat assistant answering 0G builder questions with citations, powered by 0G Compute.