Author: Satou & Shigeru

The combination of Crypto and AI Agent has Becoming one of the most compelling narratives of the moment. With the continuous iteration and innovation of technology, AI Agent is expected to become one of the tracks with the most development potential and attention in the encryption field in 2025, becoming the core driving force of this round of market. This article will sort out the current market structure of AI Agent from three levels: framework, Meme and application.

AI Agent Framework: Layer1 in the AI ​​​​field

AI Agent framework is the core technical foundation layer of AI Agent. The framework lays an important cornerstone for the development, deployment and collaboration of AI Agent. Therefore, the current competition and competition regarding the AI ​​Agent framework is actually a competition for Layer 1 in this field. At present, judging from the market value of tokens, G.A.M.E, Eliza and Swarms are in a three-legged position, and Rig and Zerepy still have opportunities to catch up.

1. G.A.M.E

G.A.M.E is a framework developed by the Virtuals team, and its design core The idea is to adopt a modular design to allow multiple subsystems to work together to jointly control the behavior, decision-making and learning process of the AI ​​Agent. These modules include "Agent Prompting Interface" as the main entrance for developers to interact with Agent behavior, "Perception Subsystem" which is responsible for processing input data and converting it into a suitable format, and "Strategic Planning Engine" which is responsible for generating specific action plans based on input information. etc. Users only need to modify the parameters of various modules to participate in Agent design. The specific modules and architecture are shown in the figure below.

The core features of G.A.M.E are:

Modular design: the entire framework is clear and easy to understand, no additional design is required;

Provides a low-code or no-code interface: greatly lowering the technical threshold.

This makes G.A.M.E particularly suitable for those who need rapid deployment , and do not care about complex technical settings. But for complex projects that require deep customization or complete control of all aspects of the Agent, G.A.M.E is not suitable.

2 . Eliza

Eliza is an open source multi-Agent framework developed by ai16z, using TypeScript as the programming language. The framework is built around a system called Agent Runtime, and its core features include:

Role system: supports simultaneous deployment and management of multiple personalized AI Agents, supported by model providers;

Memory Manager: Provides long-term memory and context-aware memory management capabilities through the Retrieval Augmentation Generation System (RAG);

Action System: Provides smooth Platform integration enables reliable connection with social media platforms such as X

Eliza revolves around an Agent. The runtime system is built to be seamlessly integrated with the character system, memory manager and action system. Eliza also supports a plug-in system with modular function expansion, which can realize multi-modal interaction such as voice, text and media, and is compatible with Llama and GPT-4. and AI models such as Claude. Therefore, Eliza is suitable for projects that require deeply customized solutions and complex cross-platform multi-agents.

3. Swarms

Swarms is an open source multi-Agent orchestration framework developed by founder Kye Gomez. Its core idea is to allow multiple AI Agents to collaborate and use collective wisdom to solve problems. Complex problems. Its core features include:

Multi-Agent collaboration: SWARMS provides a transparent and traceable environment for multiple Agents, allowing different Agents to Work together to improve task execution efficiency

Incentive mechanism: SWARMS uses tokens as an incentive tool for Agents. The system will dynamically allocate tokens based on the difficulty of the task and the quality of the final result.

Data security: SWARMS uses distributed storage and multi-party secure computing (MPC) technology to ensure that privacy and data security can be protected when exchanging data between agents.

These characteristics of Swarms enable it to give full play to its advantages in multiple complex fields and provide a high degree of reliability and scalability according to needs.

4. Rig

Rig is a Rust-based open source framework developed by the ARC team and designed to simplify the development of large language model (LLM) applications. The Rig framework has the following characteristics:

Unified interface: Provides a consistent interface, supports multiple LLM providers (such as OpenAI and Anthropic) and multiple vector storage (such as Seamless interaction between MongoDB and Neo4j).

Modular architecture: The framework adopts a modular design, including core components such as "Provider Abstraction Layer", "Vector Storage Integration" and "Agent System" to enhance System flexibility and scalability.

Type safety and efficient performance: Use the Rust language to achieve type safety, avoid compile-time errors, and improve concurrency processing capabilities through asynchronous operations. Efficient serialization and deserialization processes built into the framework optimize data processing.

Error handling and recovery: The built-in error handling mechanism improves the ability to recover from LLM service provider or database failures and ensures the stability of the framework.

These features allow different LLM models and storage backends to be easily integrated on the same platform. Therefore, Rig is suitable for developers who want to build AI applications in Rust and projects that have high requirements for performance, reliability, and security. However, the Rust language itself has learning costs.

5. ZerePy

ZerPy is an open source framework written in Python. ZerePy focuses on simplifying the development and deployment process of personalized AI Agents, especially Application scenarios for content creation on social platforms. Through this framework, developers can easily create AI that can post, reply, like and forward on social media. Agent. In addition, ZerePy is also particularly suitable for creative fields such as music, memos, NFT, and digital art. ZerePy performs well in creativity and is suitable for quickly deploying some lightweight Agents, but its application scope is relatively limited compared with other frameworks. Narrow.

The basic framework is AI The important directions of the Agent track, judging from the most popular frameworks at present, they all have different characteristics and have their own applicable scenarios, but the comprehensive goal is to build a comprehensive AI Agents ecosystem and become a large-scale application of intelligent Agents. A solid platform. In the future, as these frameworks are further improved and upgraded, they will become a springboard for the launch of various projects and a fertile ground for the growth of the value of various tokens.

AI Meme: The first successful appearance of AI Agent

Meme coins have always been an important concept segment in the crypto asset market. Different from traditional Meme coins, AI Meme is driven by AI Agent, and the culture or phenomenon behind it is presented by Agent. With GOAT, FARTCOIN As the market value of AI Meme coins continues to grow, AI Meme has also received more and more attention. It can be said that AI Meme is the first successful appearance of AI Agent in the encryption market.

1. GOAT

The project that really started the AI ​​Meme is Goatseus Maximus. This story began in March 2024, when developer Andy Ayrey launched a project called Infinite Backrooms Escape. Experimental system that integrates multiple large language models and allows them to talk to each other. The experimental results show that the dialogue between AIs shows extremely creative interactions without restrictions, and even gave birth to a project called GNOSIS. OF GOATSE's surreality. Next, Andy and Clau.de Opus co-authored a research paper on how AI creates memes, and GOATSE is analyzed as the first case. This series of explorations eventually gave birth to the AI ​​Agent "Truth of Terminal" (ToT). In July, a16z co-founder Marc Andreessen discovered ToT’s tweets and transferred $50,000 in Bitcoin to ToT’s Bitcoin wallet through a series of conversations. On October 10, an anonymous person released the GOAT meme coin on the social platform, which was publicly supported by ToT. The market value of the GOAT meme coin surged in just a few days. Andreessen's donation brought huge exposure to GOAT and became one of the key factors driving GOAT's rising market value. GOAT's highest market value has exceeded US$1.3 billion.

2. Fartcoin

The birth of Fartcoin is closely related to GOAT, both of which originated from ToT. During the Big Language Model conversation, it was mentioned that Musk likes the sound of farts and proposed the creation of a token called Fartcoin. Based on this dialogue, Fartcoin came into being, slightly later than GOAT. Fartcoin also attracted some attention with its clever timing of birth, but it was not as good as GOAT at the beginning. After that, on November 16, the number of Fartcoin’s Twitter followers suddenly doubled in just a few hours, and the price also increased by about 15%. However, this increase failed to receive widespread and sustained discussion. On December 13, Marc Andreessen retweeted a tweet about Fartcoin, but the tweet did not lead to a sharp increase in the price of the token. The main reason for the growth of Fartcoin price may be some main funds. Because among the earliest buying addresses, the investment fund Sigil Fund is suspected to have appeared. In addition, the founder of Sigil Fund has repeatedly expressed his optimism for AI Meme on Twitter, and even proactively forwarded a tweet asking whether Sigil Fund holds Fartcoin. Fartcoin eventually received widespread attention from social media, with its highest market value exceeding $1.5 billion.

AI Agent application: Agent can do more

With the further advancement of AI Agent in the field of encryptionApplications, market focus has also expanded from AI-driven pure meme coins such as GOAT and Fartcoin to more interactive and creative AI Agent applications.

1. Entertainment Agent

The first practical application of AI Agent is entertainment. Examples include Luna and the aforementioned ToT. Luna is a virtual idol that is tightly integrated with its native token LUNA and launched as part of the Virtuals platform. Luna will broadcast live on social media 24 hours a day and tweet frequently. Therefore, the quality of Luna's live broadcasts and tweets is one of the key factors affecting its market value. However, currently, Luna's token growth space under this model is limited. In contrast, ToT’s tweets mainly focus on original and humorous content. It is not tied to GOAT or other tokens. Although ToT occasionally mentions GOAT tokens, this is not its core focus. Both AI Agents, Luna and ToT, have tokens that play a key role in narrative promotion. For Luna, the token represents the core meaning of its existence, while for ToT, the GOAT token has become an important tool to expand its influence.

2 Investment Research Analysis Agent

In addition to entertainment applications, AI Agent can also be used for investment research analysis in the encryption field. Currently, the most popular Agent in this field is aixbt. aixbt is an AI Agent released on Virtuals Protocol. It focuses on analyzing hot topics and trends in the cryptocurrency market, especially discussions from social media platforms such as X, to help users quickly grasp market changes and potential investment opportunities. aixbt continues to have the highest CT user attention on Kaito, and its capabilities have tended to surpass human KOLs.

3. DeFi + AI Agent

If Luna and aixbt do not have much practical effect and are still at the Meme level, then the combination of AI Agent and DeFi truly gives Agent practical application scenarios. This combination of DeFi and AI Agent is called DeFAI. The development of DeFAI includesTwo major directions: Agent assisted users and Agent autonomous transactions.

Agent assists users

AI Agent assists users mainly to simplify the complexity of DeFi operations , allowing more ordinary users to easily participate in and manage DeFi projects. Users can use natural language to directly guide the AI ​​Agent to perform tasks, thus shielding complex technical details. There are some DeFAI projects on the market that have begun to emerge. Take Griffin and Neur as examples. Both are AI assistants built on Solana, which can help users complete wallet creation and management, token analysis, token transactions and other operations. In terms of user experience, Griffin provides users with more functions, while Neur provides relatively fewer but more detailed functions, and Neur's performance is better. It can be seen from the comparison between the two that the main focus in this field in the future will focus on the perfection of functions, user experience, cost and other issues.

Agent autonomous trading

If the main body of DeFi is still the model of Griffin and Neur Human users, then Agent autonomous transactions make AI the main body of DeFi. Unlike past trading robots that were limited to executing preset trading strategies, AI Agents are able to obtain real-time information from the market environment, perform contextual analysis, learn market trends and adjust strategies based on these data. This enables the Agent to make more precise decisions in a dynamically changing market and perform complex operations beyond the original program settings. Related projects include Cod3x, Almanak, etc., but this field is still in the preliminary development stage, and these projects have yet to be tested by the market. There is no doubt that the biggest obstacle to Agent's autonomous trading is the issue of trust. First, it is necessary to trust that the relevant operations are actually performed by the Agent. Second, it is necessary to trust that the Agent's trading strategy will not lead to unnecessary losses. Future projects must address these trust issues if they are to make a difference.

After several months of development, AI Agent in the encryption field has gone from pure meme to entertainment application, and then to several stages of practical application. In fact, encryption practitioners have never stopped exploring the possibility of Crypto x AI. Since 2023, CGV Research has continued to pay attention to the project progress of the Crypto x AI track.

In the future, as the infrastructure continues to mature, the Agent system becomes more intelligent and stable, and anyone can easily deploy and use Agent through natural language. At this time, the Agent framework will become an infrastructure, and other Various applications will be built based on these frameworks. The valuation of the Agent framework is expected to continue to see breakthroughs, and some Agent application projects may further capture market attention and investment value due to their excellent business capabilities and user experience.