DGrid AI functions as an AI infrastructure initiative that links the need for model inference with model suppliers and distributed node runners via a single gateway, marketplace, and blockchain coordination system. It aims to enable developers to reach various AI models through one interface, while permitting providers to list services and earn payments based on usage.

The DGrid AI Gateway provides an API compatible with OpenAI standards, granting access to over 200 models covering text, reasoning, coding, image, and multimodal tasks. Instead of managing separate integrations for every provider, developers can utilize one endpoint and authentication process. Features include routing, provider fallback options, usage tracking, cost management, and tools for selecting models. The ecosystem also features an AI Arena for comparing model outputs and a Model Marketplace where providers list models and set prices.

The decentralized network structure defines roles for users, developers, model providers, and node operators. Inference requests are directed to available providers, with verification and settlement mechanisms intended to log service delivery and manage payments. The project’s Proof of Quality framework aims to assess both model responses and node performance. Additional tools include Dori, an assistant for model selection, and DClaw, a layer for deploying persistent AI agents controlled by users.

DGAI serves as the native token for the DGrid ecosystem. Its planned uses include staking by node operators and service providers, paying for AI inference and agent services, distributing network and ecosystem rewards, and taking part in protocol governance. The token follows the BEP-20 standard on BNB Smart Chain and is also available on Arbitrum One for cross-chain functionality. The total maximum supply is capped at 1,000,000,000 DGAI, with no plans for inflationary minting after launch.

The system also supports x402-style per-request payment authorization on BNB Smart Chain. This allows applications to attach payment authorization to inference calls and receive either streaming or standard JSON responses. The goal is to link AI usage with programmable payments while maintaining compatibility with standard developer tools. Staked DGAI may serve as collateral for service accountability, and network rules could penalize nodes that are malicious or unavailable.

Token distribution allocates 50% to node and infrastructure incentives, 15% to community programs, 10% to team incentives, 10% to investors, 8% to airdrops, and 7% to initial liquidity. Node incentives will be released over ten years, with emission cuts occurring every two years. Allocations for the community, team, and investors face lock-up periods and linear release schedules, whereas airdrop and initial-liquidity portions unlock at the token generation event. Separate details on reserve addresses and release schedules are provided for verification.

The live DGrid AI ecosystem comprises the AI Gateway, AI Arena, Model Marketplace, developer documentation, and integration guides for popular AI development tools. The project shares its contracts and related code via its GitHub organization and offers public documentation detailing token functions, allocation, circulation, and network components.