bitpro가 원문에서 글자만 모은 내용이에요. 표·그림·수식은 빠지거나 깨질 수 있어요. 전체 3쪽 중 2쪽
ide a solid foundation for an intelligent, trustworthy,
and self-driven digital future. The design follows a layered, modular approach to ensure
scalability, flexibility, and security. From the bottom up, the architecture comprises four layers:
the Infrastructure Layer, Core Components Layer, Intelligent Services Layer, and Ecosystem
Interaction Layer.
1. Infrastructure Layer
This layer serves as the ecosystem's trust layer and computational base, ensuring a
decentralized, secure, and reliable environment.
Blockchain-Based Settlement
Built on the TRON network, this layer issues assets, settles transactions, and runs governance
smart contracts. It benefits from TRON's high throughput, low fees, and active ecosystem.
11
Distributed Compute
Decentralized computing power is pooled to handle heavy AI training and inference. This avoids
single points of failure and ensures no single cloud can monopolize the data.
Decentralized Storage
BTFS is used as a decentralized storage solution to keep AI model parameters, key interaction
data, and metadata immutable, traceable, and available in the long term.
2. Core Components Layer
This layer supplies the core intelligent modules and the economic engine for building
AI-powered DApps.
Multi-Agent Framework
Agent runtime environment: A secure sandbox for deploying and running AI agents. Developers
can run agents on local machines, in the cloud, on edge devices, or mix these options as
needed.
Communication and collaboration bus: A message-passing and event-driven hub that allows
agents to communicate with other agents and smart contracts in a secure, asynchronous
manner. It supports complex multi-agent workflows where different agents can coordinate and
collaborate.
Unified API suite: Provides a complete set of developer-friendly interfaces, ranging from
lower-level Core APIs to higher-level AgentChat APIs. The Extensions API and Magentic API
are used to plug in external tools such as LLM calls, code execution, and web browsing,
lowering the learning curve and speeding up delivery.
Asset Management Core
Visual agent minting: With a minimal-code/no-code visual interface, users can set an AI agent's
behavior, capabilities, and economic rules in a simple graphical flow, then mint the matching
AINFT and native token with one click.
Economic engine: A built-in, configurable smart-contract library for tokenomics that integrates
the models of SunPump and NFT Pump. It automates token issuance, sets up liquidity pools,
allocates fees, and rewards the community so each AINFT runs its own sustainable economy.
12
3. Intelligent Services Layer
This layer turns AI capabilities into ready-to-use intelligent services for upper-layer apps.
DeFAI Services Module
On-chain strategy library: Offers a set of audited, composable DeFi smart-contract templates
that cover AMM optimization, cross-market arbitrage, and lending strategies. AI agents can call
and execute these contracts automatically.
Off-chain data oracles and AI oracles: In addition to securing price feeds from external sources,
the system introduces AI oracles that deliver validated insights to the chain, including sentiment
analysis and risk assessments. Smart contracts can then make better-informed decisions on
complex issues.
Decentralized AI Model Services: AINFT Grid
Collaborative training framework: Data owners can join model training by contributing datasets
or computational power while keeping raw data private. Rewards are paid in tokens based on
each participant's measured contribution.
Model marketplace and inference services: Trained models can be packaged as AINFTs, then
listed for trading or licensing in a decentralized marketplace. Users pay to run inference, and
revenue is split automatically among model contributors according to preset rules.
4. Ecosystem Interaction Layer
This layer serves as the bridge connecting the ecosystem, the wider world, and users.
Social Media Compatible
Social media adapters: Standard plugins and APIs enable AINFT agents to connect securely to
Twitter, Discord, Telegram, and other major platforms. Agents can publish content, handle
community interactions, and run user-growth campaigns on their own.
AI interfaces: Users interact with their AINFT agents through the web, mobile apps, or
immersive 3D spaces. They can type, speak, or use simple visual controls, making the
experience straightforward and intuitive.
13
Developer and Community Hub
Comprehensive developer suite with SDKs and APIs: Offers detailed documentation, working
code examples, a testnet environment, and developer tools. This makes it easy for third-party
developers to build new apps and agents on the AINFT stack.
User dashboard: A single home where token holders propose and vote on core protocol
upgrades, treasury spending, and ecosystem fund allocations, enabling transparent,
community-led governance.
V. Business Ecosystem
After years of development, AINFT's vision is taking shape through a suite of business
initiatives.
1. AINFT Agent Framework—Multi-Agent System Framework
AINFT Agent Framework is a multi-agent system (MAS) framework created by the AINFT team.
It is designed to provide developers with a complete set of components and toolchains for
building AI agents, with a focus on multi-agent collaborative workflows. The framework adopts a
layered, modular, and highly extensible architecture, where each layer has clearly defined
responsibilities and low coupling. This ensures the stability of core functions while offering
flexibility for custom extensions.
The AINFT Agent Framework offers the following key API modules as part of its core
capabilities:
Core API: Delivers a message-passing communication system together with an event-driven
model for scheduling agents. It allows agents to operate seamlessly across local and distributed
environments and serves as the backbone for orchestrating complex MAS workflows.
AgentChat API: Streamlines agent interactions through a simple, high-level interface. Intuitive
and fully featured, it flattens the learning curve and is perfectly suited for building quick
prototypes and testing new ideas.
Extensions API: Offers extensibility through plug-in enhancements ranging from LLM clients to
dynamic code execution, giving developers the freedom to tailor agent behavior.
14
Magentic API: Processes external resources and runs tasks. It provides simple, standard
interfaces for web data crawling, code execution, and reading and writing files, expanding the
real-world use cases for AI agents.
With this design, the AINFT Agent Framework streamlines multi-agent development while
balancing performance, security, and maintainability. It lays a strong foundation for the next
generation of distributed AI applications.
2. AINFT Nova—AI Agent Launch Platform
AINFT Nova is a groundbreaking platform for designing AI agents and launching them as
tokens. Simplifying both agent creation and tokenization, it sets the stage for deeper fusion of AI
and blockchain and for innovative use cases. The platform enables users to quickly deploy AI
agents with personalized interaction and decision-making skills, and to issue a dedicated token
native to each agent at the same time. This brings agent behavior and economic incentives
together, creating a self-sustaining agent economy.
In terms of core functionality, the platform offers an intuitive, visual interface for configuring AI
agents. Through graphical tools that require minimal coding, users can easily build agents
capable of natural language processing, emotion recognition, and advanced social interaction
without needing any programming expertise. This makes development significantly faster and
more accessible.
On the tokenomics side, AINFT Nova integrates the models of SunPump and NFT Pump,
automating token issuance, liquidity pool creation, and revenue distribution end to end. Each AI
agent runs its own economy and community incentives, reflecting the principle of one agent,
one economy, and keeping value circulating across the ecosystem.
On the operations and engagement front, the platform supports automated social media
management across multiple platforms. The AI Agent publishes posts, answers comments, and
connects with followers across platforms such as X. It learns continuously and adjusts its style,
making interactions more natural and genuine and boosting user engagement.
Additionally, the system is built for scale. It supports flexible integration with a range of social
platforms, including Discord and Telegram, and it is cross-chain compatible. In the future,
support will grow to cover multi-chain and omni-chain environments, unlocking more scenarios
and applications for AI agents.
15
Looking ahead, AINFT aims to support the integration of external multimodal foundation models
— including text-to-video systems like Sora — into its decentralized AI ecosystem. This will
allow developers to plug in state-of-the-art generation capabilities directly into AINFT agents,
enabling seamless creation, tokenization, and monetization of AI-generated media within a fully
on-chain environment.
With AINFT Nova, creators, developers, and teams can easily launch AI agents that combine
social skills with built-in economics. This opens the door to a large-scale AI agent economy and
drives AI and blockchain toward a more open, empowering, and innovative future.
3. AINFT AgenTX—FinAgent Trading Framework
AINFT AgenTX (FinAgent Trading Framework) is a next-generation AI-driven trading
infrastructure for DeFi. By deeply integrating AI with blockchain technology, it delivers trading
solutions that are intelligent and secure, with the capacity to continuously evolve. Built on a
modular, scalable system, the framework integrates a wide range of AI models: machine
learning models for market forecasts, reinforcement learning models for optimizing strategies,
and NLP models for context-aware decision making. With this stack, developers can efficiently
design, backtest, and deploy advanced quantitative strategies, especially for medium to low
frequency trading in multi-chain DeFi, helping grow assets and manage risk across changing
market conditions. In addition, all AI agents support the Agent Commerce Protocol (ACP) so
assets can be managed interactively across the platform. This makes trading broader in reach
and deeper in strategy.
AINFT AgenTX handles asset management through smart contracts, delivering complete
on-chain custody and automated execution. Smart contracts handle all asset operations and
strategy scheduling, making money flows and strategy execution tamper-proof and easy to
trace. Key mechanism upgrades and system parameter adjustments are governed by
DAO-based community decisions, independent of centralized control. All major changes must
be approved through a vote by token holders. This governance model ensures both fairness
and transparency while effectively mitigating single points of failure and malicious manipulation
risks, thereby providing institutional-grade security for user assets.
AINFT AgenTX embraces open-source development and collaboration at its core. The platform
makes its core protocols and algorithms fully available, accompanied by comprehensive
technical documentation, an extensive set of API interfaces, and a range of practical example
projects. These resources cover the entire development cycle—from strategy design and
backtesting to live deployment. The framework encourages community-driven innovation by
regularly hosting hackathons, offering ecosystem incentive programs, and providing open
access to governance participation. These efforts invite developers and research teams to push
the boundaries of algorithmic trading, shape industry standards, and create a more inclusive
and innovative DeFi trading ecosystem.
16
AINFT AgenTX aims to be the key infrastructure connecting AI with DeFi. With its highly
autonomous, composable, and transparent algorithmic trading framework, the platform lowers
the technical barriers and operational risks of participating in the DeFi market, enabling users to
engage with the evolving digital finance landscape in a more intelligent, efficient, and accessible
manner.
4. AINFT Grid—Decentralized AI Model Platform
AINFT Grid is an open platform designed to drive decentralized AI model training and
deployment. By integrating distributed computing with blockchain technology, it aims to build a
more equitable, transparent, and efficient environment for AI development. Through a
decentralized, distributed AI training framework, the platform transcends the limits of centralized
systems, integrating global computing resources to lower costs and maximize efficiency.
AINFT Grid specializes in the foundational creation of LLMs and DeFAI models through
collaborative training across distributed computing nodes. This approach strengthens model
robustness and generalization while unlocking new levels of innovation and real-world utility. By
leveraging smart contracts for task scheduling and incentive distribution, the platform fosters a
sustainable, self-evolving ecosystem that fuels the creation and adoption of next-generation AI
models.
Committed to open-source principles, AINFT Grid provides full transparency into its training
architecture, core algorithms, and data processing workflows. It provides detailed,
developer-friendly documentation and interface specifications that span the entire model
lifecycle—from training and fine-tuning to deployment. Through clear technical guidance and
diverse example projects, the platform significantly reduces barriers to entry for developers
participating in decentralized AI training, fostering collaboration and knowledge sharing while
accelerating technological iteration and ecosystem innovation.
AINFT Grid represents more than a technical platform—it serves as critical infrastructure for
advancing toward an open AI ecosystem. The platform is committed to democratizing AI model
training and distribution, empowering users and developers with greater autonomy while
maximizing the social value and innovation potential of AI—without compromising data privacy.
17
VI. Roadmap
1. TRON Ecosystem Intelligence
TRON Transaction Booter: An LLM-powered system for executing a wide range of transactions
on the TRON network.
TRON Data Searcher: A multimodal LLM-based system for querying diverse data across the
TRON network.
TRON Program Assister: A foundation model-powered assistant that provides smart contract
coding assistance and automated vulnerability detection.
2. TRON AI Agent Platform
Positioning TRON as a decentralized infrastructure for AI agent deployment and interaction.
AI Agent Layer: A Decentralized Identity (DID)-based framework for AI agent registration,
interaction, and execution.
SDK Layer: A software development kit that enables streamlined creation of AI agents.
User Interaction Layer: An intuitive interface for discovering agents, assigning tasks, and settling
fees.
18
Incentive Layer: An economic model designed to incentivize AI agent developers.
AI Agent Commerce Protocol (ACP): A universal protocol enabling commercial interactions and
value exchange between AI agents within the ecosystem.
3. TRON DeFi Intelligence
Delivering AI-powered decision support and quantitative tools to build a smarter DeFi
ecosystem.
Project Discovery: AI-powered automated identification of high-potential projects and protocols.
Quantitative Trading Platform: AI-assisted investment decision-making, strategy design,
evaluation, and on-chain deployment.
4. Foundation Model Decentralization
By leveraging blockchain, the platform addresses the high cost of training foundation models. AI
should be positioned as a public good countering the current trend toward closed, proprietary
systems. By allowing token holders to share in the value they generate, the platform promotes
models that are more open, decentralized, and accessible to all.
19
VII. AINFT Governance Token—NFT
1. Definition
AINFT exclusively uses NFTs as its official governance token within the ecosystem. Holders of
the token can participate in AINFT governance.
2. Allocation
Token name: NFT
Underlying blockchain: TRON
Protocol standard: ERC20/TRC20
Issuance date: May 20, 2021
Issuance platforms: Leading crypto exchanges such as HTX
Total supply: 999,990,000,000,000 tokens
Initial issuing price: 0.000012 USD
Token allocation: artist partners—30%; DeFi airdrop and mining—19%; AINFT team—19%;
initial exchange offering—2%; NFT work purchases—20%; partnerships—10%.
*All NFT tokens were fully unlocked as of May 1, 2023.
3. Business Model
NFT holders can tap into AINFT's full suite of products and unlock tiered perks based on their
holdings. By staking NFTs in the AINFT Foundation, they earn voting rights, empowering them
to join decentralized governance, from deciding on artwork within the AINFT DAO ecosystem to
participating in different levels of AINFT events.
4. Reward Mechanism
(1) DeFi Airdrop and Mining (Concluded)
Previously, users could earn NFTs by participating in liquidity airdrops and mining campaigns for
BTC, ETH, DOGE, TRX, BTT, JST, SUN, and WIN. Platforms supporting these activities
included SunSwap, JustLend DAO, and SUN.io, among others.
20
(2) NFT Airdrop and Mining (Concluded)
NFTs were distributed solely through decentralized airdrops and mining programs; holders of
BTC, ETH, DOGE, TRX, BTT, JST, SUN, and WIN received NFTs as rewards.
(3) Governance Rewards
Participants who engage in governance activities within the AINFT ecosystem can earn
additional NFT rewards.
(4) Campaign Rewards
Users may also receive NFT rewards for taking part in special campaigns and events across the
AINFT ecosystem.
VIII. AINFT Team
1. AINFT Chair: Steve Liu
The chair of AINFT, Steve Z. Liu, holds a bachelor's degree in Electrical Engineering from
Zhejiang University and an MBA from Columbia Business School. With over 20 years of
experience in global capital markets, Liu has held senior roles at leading financial institutions
including Fidelity International, Salomon Smith Barney, Nomura Holdings, and Ant Financial. He
also successfully established and managed institutional equity business for a number of leading
Chinese securities firms.
2. Partner Artists
AINFT plans to invite influential art creators and researchers from around the world to join the
AINFT advisory team. We hope to build a cutting-edge stage for art exchange to encourage
creativity, drive the value of artworks and lead the trend in the future art industry.
21
3. Other Members
Most of the AINFT team members are experts from the traditional art industry with experience in
renowned art institutions, such as Christie's and Sotheby's. The technical backbone is formed
by veteran engineers with prior experience at top internet companies like Baidu, Alibaba, and
Tencent, with expertise in blockchain and cryptography. The team's breadth and depth of talent
rank it among the industry's leaders.
IX. Resources & Contacts
AINFT official website: http://ainft.com
AINFT Marketplace: http://apenft.io
NFT Pump: http://nftpump.meme
AINFT Foundation: https://foundation.apenft.io/
Email: contact@ainft.com
TRON blockchain explorer: https://tronscan.org
22
Appendix: The AINFT Art Collection
1. Pablo Picasso: Femme nue couchée au collier
The year 1932 marked a pivotal moment in Pablo Picasso's life. In the first half of that year, he
created 프로젝트로 돌아가기