Compiler: Blockchain in Vernacular


People are already tired of AI response bots, and this type of application is now somewhat saturated. Right now, all eyes are on the next wave of truly practical AI agents.

The next wave has arrived: DeFAI (also known as AiFi or AgentFi). In simple terms, we are leveraging AI agents to simplify decentralized finance (DeFi), making it more profitable for us. The reasons I am optimistic about this field are:
• Abstraction: It addresses the complexity of DeFi interfaces through a natural language interface, making the user experience friendlier.
• User Profitability: By automating trading, optimizing yields, and using AI-enhanced trading strategies, it helps users gain more profit and reduces costly human errors.
• Yield Potential: There are already some protocols, such as Bananagun and Photon, that make trading simpler; they can almost 'print money'. Once product-market fit is achieved, we could see some interesting token economic models.

Finally... the market atmosphere has begun to shift. You can feel that everyone's focus is turning to DeFAI. I believe that in the coming months, billions of dollars will flow into this field, and now is a great time to position yourself (if you pay attention, even today, some projects have already broken their all-time highs despite the market downturn).

1. Noteworthy leading protocols (ranked by market cap)

1) Griffain (@Griffaindotcom - Market Cap: $494 million)

Created by veteran Solana developer @tonyplasencia3, who previously created @backanunderdog and @blinkdotfun. Griffain allows you to search and complete any on-chain operation through natural language commands.
You can think of it as an App Store, using different agents to complete various tasks. Some task demonstrations I have seen include: follow trading, creating memecoins, and even automatically locking in profits for you.
2) Hey Anon (@HeyAnonai - Market Cap: $172 million) Recently launched on the Sonic platform, with price fluctuations that are quite astonishing. It focuses on natural language trading interfaces, automated DeFi agents, and research and communication agents. @DWFlabs has established an AI agent fund and has already invested $20 million into Hey Anon. Some demo recordings have already showcased its powerful features, and public testing will open at the end of January. So, what has suddenly made Anon so popular? The answer is founder Daniele Sesta, who created two billion-dollar protocols (Spell and Wonderland) in the last cycle. Can he create a third one this cycle?
3) Mode (@ModeNetwork - Market Cap: $117 million) This is an Ethereum Layer 2-based protocol dedicated to building a DeFi economy fully operated by AI agents. Currently, there are 129 AI agents on the Mode network, which have completed over 1,600 DeFi transactions, although some of these agents have not yet issued tokens.

  • Arma by Giza: An agent optimizing USDC yield farming, suitable for various applications on the Mode platform.

  • Brian: Helps users execute trades and deploy contracts through natural language operations, simplifying processes.

  • Sturdy: Utilizing AI-driven yield Vaults to aggregate and optimize returns.

  • MODIUS by Olas: Allowing you to co-own or initiate your own AI agents, optimizing liquidity mining strategies.

  • Amplifi: Helping you achieve the best returns across multiple chains through cross-chain liquidity and AI optimization.

4) Orbit (GRIFT Market Cap: $43 million)

Orbit's agents allow users to interact with different protocols across multiple chains. It was built by the SphereOne team, a crypto payment platform that raised $2.5 million in 2023. Orbit has already supported over 100 chains, including EVM, Solana, Aptos, Sui, Cosmos chains, and Bitcoin.
SphereOne aims to create a complete onboarding process that simplifies all steps from purchasing cryptocurrency to executing on-chain operations through AI agents.
5) Neur (Market Cap: $41 million) Recently, Binance Alpha listed Neur. Neur enables users to interact with major Solana-native protocols through natural language. It is a fully open-source, community-driven project, and the team regularly launches reward programs to encourage developer contributions. Since its launch, activity on GitHub has been steadily increasing, currently boasting 350 stars.
The product is still in testing, and users need to pay 1 $SOL to join its limited early access program.
6) ASYM (@ASYM41b07 Market Cap: $21 million) This is an AI-driven trading agent that has demonstrated 3-4 times return rates in real trading, such as $NEUR and $GRAPH. Profits will flow back to $ASYM, driving buying pressure, and 28% of the supply has already been locked. ASYM focuses on DeFi automation and agent infrastructure, capturing memecoin trends using on-chain data and social data, striving to consistently outperform in the market.
7) Gekko AI (@Gekko_Agent Market Cap: $16 million) A 99% fair launch trading agent created by @getaxal. Although the Axal team does not hold any tokens, Gekko, as a trading agent, offers bold crypto strategies and has integrated with Axal's Autopilot system, driving traffic and optimizing products.
8) Axal @getaxal Axal @getaxal is a verified agent network backed by a16z and Unusual Ventures. Axal is currently in closed testing, aiming to apply autonomous agent technology to everyday tasks, from trading to flight bookings.
2. Upcoming noteworthy tokenless projects 1) Cod3x @Cod3xOrg Cod3x is a DeFAI ecosystem that allows users to create trading agents without code. Anyone can use the growing strategy and API library to build their own trading agents. It integrates machine learning models from @AlloraNetwork to support more complex strategies and uses cdxUSD liquidity to ensure stability. Its flagship agent @BigTonyXBT (Market Cap: $9 million) uses AI-driven predictions for mainstream trading. Cod3x aims to create the world's largest economic AI agent community.
2) Almanak @Almanak__ Almanak is a DeFi platform that provides institutional-level tools for building autonomous financial agents. It conducts real strategy testing by forking EVM chains and considers complex factors such as MEV and Gas. It also uses Trusted Execution Environments (TEEs) to protect privacy and handles non-custodial funds through the Almanak wallet. The project has raised $1 million on the Legion Launchpad and is expected to launch a Beta version of the agent in Q2 2025.
3) WayFinder Protocol @AIWayfinder Imagine being able to deploy your own AI agent with just a text message. WayFinder's 'navigation path' allows these agents to operate smoothly in DeFi. Token swaps, yield farming, NFT launches, and more can all be completed through short commands. It supports multiple chains, including Ethereum, Solana, and Base. Its highlight is that agents can learn from one another, meaning the success of one agent unlocks more potential for all agents. Currently, tokens have not been launched, but $Prime stakers have received 40% of the $prompt supply.
4) Slate @slate_ceo Slate is an AI tool that automates trade execution by combining real-time aggregation alerts from Telegram, Discord, X, and on-chain channels. It not only provides notifications but also executes trades instantly on Solana, Base, and Hyperliquid, equipped with MEV protection and optimized routing. Designed for traders who need precision and speed, it requires no manual operation.
3. Summary


I have been involved in DeFi since the end of 2020.

I love the sense of infinite possibilities that era brought, with new tokenomics designs and innovations emerging every week. But as DeFi matures, the pace of innovation has slowed, and bringing in AI will lead to a new renaissance for DeFi.

The market is currently quite volatile, and there's no need to rush to follow trends. We will gradually gain more public access to see if these protocols can deliver on their promises. However, this is a field worth keeping an eye on.

Some other thoughts:
• An interesting point is whether any established DeFi giants will start launching AI-based front ends.
• We are still in the very early stages, and everyone is 'selling dreams'. There are significant risks in this, as the challenges of execution are also high. Several months from now, many protocols will open testing, at which point we will evaluate them based on execution results.
• You can use these tokens to create your own DeFAI watchlist (for example, on DexScreener).

Note: If I missed any interesting protocols, feel free to comment and supplement. This field is still in a very early stage, and some protocols are still in testing, facing execution risks. So please proceed with caution.