Pundi AI is a decentralised protocol that integrates artificial intelligence (AI) development with blockchain infrastructure. It focuses on building an open ecosystem for AI agents, data generation, and value attribution, aiming to decentralise how AI is trained, deployed, and monetised. The project emerged as an evolution of Function X (FX), introducing a new design layer—Pundi AIFX—that combines decentralised infrastructure, community-led governance, and incentive-based data ecosystems.

Pundi AI addresses one of the key bottlenecks in AI development: access to high-quality, continuously updated datasets. It provides tools and mechanisms for users to generate and label data, train and deploy AI agents, and engage in governance through token-weighted proposals. The protocol includes several key components:

• Pundi AIFX Omnilayer: A hybrid infrastructure combining f(x)Core and Pundi X Chain, offering EVM compatibility, AI workload processing, and decentralised compute coordination.

• AI Agent Framework: A standardised format for creating and launching decentralised AI agents. Developers can deploy agents to perform tasks such as language generation, trading, moderation, and research.

• Pundi Fun Launchpad: A launchpad for AI agents to introduce native agent tokens. This includes an automated market-making system and bonding curve for liquidity provisioning.

• PURSE+ Browser Plugin: A tool that enables users to tag content directly from the internet, which contributes to AI training datasets. This includes structured and unstructured data collection via web activity.

• Pundi AI Data Platform: A decentralised application that allows contributors to participate in data generation, annotation, and review. It supports dataset curation with community feedback, attribution scoring, and incentive payouts.

PUNDIAI is the native utility and governance token of the Pundi AI ecosystem. It is designed to power network functions, coordinate value exchange, and ensure decentralised governance. The token was created following a 100:1 redenomination from the former FX token and serves several roles across the platform:

• Reward Distribution: Contributors who generate or label training data are compensated in PUNDIAI. This applies to users of the Data Platform, PURSE+ plugin, and participants in on-chain model training.

• Governance Participation: PUNDIAI holders vote on proposals relating to protocol changes, incentive adjustments, and AI agent launches. Voting power is proportional to token holdings, and proposals include upgrades to the Pundi AIFX layer and funding allocations.

• Service Fees: Fees on agent deployments, model access, and decentralised compute are paid in PUNDIAI.

• Token Launch Support: New AI agents can use PUNDIAI to bootstrap their own sub-tokens through the Pundi Fun Launchpad. This includes initial bonding, rewards emission, and price discovery mechanisms.

• Ecosystem Incentives: Allocations of PUNDIAI are used to support strategic initiatives, grant programmes, and liquidity provisioning for ecosystem dApps and infrastructure.