DGrid AI

DGrid AI is an AI infrastructure project that connects demand for model inference with model providers and distributed node operators through a unified gateway, marketplace, and blockchain-based coordination layer. The project is designed to let developers access multiple AI models through a common interface while allowing providers to publish services and receive usage-based payments.

The DGrid AI Gateway offers an OpenAI-compatible API for accessing more than 200 text, reasoning, coding, image, and multimodal models. Developers can use a single endpoint and authentication flow rather than maintaining separate integrations for each provider. The gateway includes routing, provider fallback, usage records, cost controls, and model selection tools. DGrid also operates an AI Arena where users compare model responses, and a Model Marketplace where model providers can publish available models and define pricing.

DGrid's decentralized network design assigns roles to users, developers, model providers, and node operators. Inference requests are routed to available providers, while verification and settlement components are intended to record service delivery and coordinate payments. The project's Proof of Quality framework is designed to evaluate model responses and node performance. Supporting products include Dori, a model-selection assistant, and DClaw, a deployment layer for persistent user-controlled AI agents.

DGAI is the native token used within the DGrid ecosystem. Its intended functions include staking by node operators and service providers, payment for AI inference and agent services, distribution of network and ecosystem rewards, and participation in protocol governance. The token uses the BEP-20 standard on BNB Smart Chain and is also represented on Arbitrum One for cross-chain use. The aggregate maximum supply is fixed at 1,000,000,000 DGAI, with no inflationary minting planned after launch.

The ecosystem also supports x402-style per-request payment authorization on BNB Smart Chain. In this flow, applications can attach payment authorization to inference calls and receive streaming or standard JSON responses. This is intended to connect AI usage with programmable payments while preserving compatibility with familiar developer tooling. Staked DGAI can be used as collateral for service accountability, and network rules may penalize malicious or unavailable nodes.

The token allocation assigns 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 are released over ten years with emission reductions every two years. Community, team, and investor allocations are subject to lock-up and linear release schedules, while airdrop and initial-liquidity allocations are unlocked at the token generation event. Detailed reserve addresses and release schedules are supplied separately for verification.

DGrid AI's live ecosystem includes the AI Gateway, AI Arena, Model Marketplace, developer documentation, and integration guides for commonly used AI development tools. The project publishes its contracts and related code through its GitHub organization and provides public documentation covering token functions, allocation, circulation, and network components.