Solana x402 Hackathon Concludes: A Quick Look at Five Major Innovative Projects

Original: Odaily Odaily Daily

Author: jk

The Solana x402 hackathon, which lasted for two weeks, successfully concluded in November, with the organizers officially announcing the winners of the main track on November 25. This remote hackathon attracted enthusiastic participation from developers worldwide, ultimately receiving over 400 project submissions. The previously popular AI payment protocol x402 is an internet-native payment protocol developed by Coinbase, aimed at enabling AI programs to autonomously complete online payments like humans. The vision is to have your AI assistant not only help you look up information but also pay for data and subscribe to services on its own, all automatically completed on the blockchain.

This hackathon has established five competition categories, each with a top prize of $20,000. Now, let's take a look at the innovations of these five award-winning projects brought to you by Odaily.

Intelligence Cubed (i³): Let AI models trade like stocks

Intelligence Cubed has created a very interesting platform that can be understood as “Taobao + stock market for AI models”. On this platform, AI models can not only be used but also bought, sold, and invested in.

Imagine a scenario like this: you are a developer of an AI model who has spent a lot of time training a powerful image recognition model. In the traditional model, you might need to set up your own server, handle payments, and manage users. But on the i³ platform, you only need to upload the model, set the price for each call (for example, 0.01 dollars), and the platform will automatically handle everything.

Interestingly, i³ introduces the concept of “model tokenization.” Developers can divide the ownership of a model into multiple shares and sell them through IMO (Initial Model Offering, similar to a stock IPO). After investors purchase model tokens, whenever someone uses the model and pays a fee, the token holders will receive a proportional share of the revenue. If someone creates an improved version based on your model, your original model will automatically receive “royalties.” The project also introduces the concept of “open source threshold,” where the model will automatically turn open source when more than 51% of its ownership is held by the public, to accelerate adoption and re-creation.

Technically, i³ deeply integrates the x402 payment protocol. Each time a user wants to invoke an AI model, the system first generates a payment request, showing how much USDC needs to be paid. After the user confirms the payment through the Phantom wallet, the transaction is verified on the Solana blockchain, and the entire process only takes a few seconds. The AI model will only start working and return results after the payment is confirmed. The platform also provides a visual workflow editor, allowing users to connect multiple AI models like building blocks to create complex processing flows, with clear costs for each step.

PlaiPin (Solana ESP32 Native x402): Let IoT devices learn to spend money on their own.

What PlaiPin is doing sounds a bit futuristic: they have enabled a microchip (ESP32) that only costs a few dollars to manage its own wallet and make payments on its own. What does this mean?

Imagine you have a smart temperature sensor that collects data every day. In the traditional model, this sensor needs to send the data to a cloud server, where humans decide whether to sell the data. But with this technology, the sensor itself can become an independent “merchant”: it can determine when the data is valuable, contact buyers on its own, collect payments, and then store the money in its own blockchain wallet.

For example, your smart refrigerator detects that it needs to call an AI service to optimize the temperature control algorithm, and it can pay $0.001 to purchase this service without any human intervention. Or your vacuum cleaner encounters complex terrain while cleaning and needs to purchase a one-time advanced navigation algorithm call, and it can also complete the payment autonomously.

Technically, the breakthrough of this project lies in packing a complete blockchain wallet and payment capabilities into a small chip. The ESP32 chip stores its own keys internally (similar to a bank card password) and can perform cryptographic signatures to prove “this money is indeed what I want to pay.” The entire payment process takes about 2-4 seconds: the device detects the need for a paid service, automatically parses the price and payment address, signs the transaction internally on the chip, submits it to the blockchain network through a facilitator (which can be understood as a payment channel), and finally receives the service. The key point is that the user's wallet private key never leaves the chip, ensuring security.

The project code has been tested successfully on real hardware, and the developers have provided detailed installation guides, allowing anyone to purchase a set of hardware for just a few dozen dollars to try it out. This opens up a whole new business model for IoT devices: turning them into “electronic life forms” that can actively participate in economic activities.

x402 Shopify Commerce: Let your Taobao store accept AI customers in 2 minutes.

If the previous projects were more technical, the x402 Shopify Commerce project is very down-to-earth. The problem it aims to solve is: how can ordinary online stores serve AI customers?

Online stores today are designed for humans: they have images, shopping carts, and checkout buttons. But AI programs “do not understand” these. This project is like installing an “AI-specific channel” for online stores: the store owner only needs to do three things – first, paste the URL and authorization code of their Shopify store (30 seconds); second, select which products allow AI to purchase (60 seconds); third, open the monitoring panel to view orders placed by AI (30 seconds). The entire process does not require writing a single line of code.

Once the setup is complete, the AI program can shop like a human. For example, if an AI assistant from a certain company receives the task “order 100 signing pens for the office”, it will automatically search your store, check the product catalog, select the appropriate items, calculate the total price, and then pay with USDC. The entire process follows the standard x402 protocol: the AI initiates the purchase request, your store automatically tells the AI “you need to pay X dollars USDC to this address”, the AI completes the transfer, and once the store verifies the receipt, it automatically creates the order. The order will appear in your Shopify backend like a regular order, and you just need to ship it as per the normal process.

This project cleverly combines two open standards: MCP (Model Context Protocol) allows AI to “understand” what products your store has, and x402 standardizes and automates the payment process. More importantly, because it uses direct transfers via blockchain, store owners do not need to pay credit card fees (usually 3-5%), and the funds can arrive within seconds.

For early-stage AI startups, this means they can purchase resources for their AI products directly from suppliers without the need for manual approval or pre-recharge. For e-commerce sellers, this opens up a whole new customer base - those AI agents who independently procure on behalf of companies or individuals.

Amiko Marketplace: Establishing Credit Profiles for AI

When AI programs start spending money to buy services, a question arises: how do I know if this AI is reliable? Will it pay for the service and then run away? Is the quality of the services it provides good? Amiko Marketplace is here to solve this problem by establishing a “credit profile” for each AI on the blockchain.

The operation of this system is quite ingenious. Every time an AI program receives its first payment, the system automatically creates an identity profile for it, recording its wallet address and basic information. Each time the AI completes a job and receives payment, the system creates a permanent work record that includes information such as who the client is, how much was paid, the transaction hash, etc. After using the service, clients can rate the AI (1-5 stars) and leave a review.

The most interesting aspect is its rating mechanism: it does not simply take the average score, but rather “weights it by payment amount”. Suppose an AI receives a 5-star rating for a $100 transaction and a 3-star rating for a $10 transaction, then its overall score would be closer to 5 stars because the evaluation weight of the larger transaction is higher. The benefit of this design is to prevent score manipulation – if someone wants to boost their ratings through a large number of small transactions, the cost would be very high, and the effect would be limited.

Take a practical example: You developed an AI translation service, and initially, there are no reviews. A customer spends $50 using your service, feels very satisfied and gives 5 stars, and now your profile has its first positive review and a record of “Total Transaction Amount $50”. As more customers use and review your service, your credit score will increase. Other potential customers, seeing that you have over 100 positive reviews and a total transaction amount of $10,000, will naturally be more willing to choose your service.

This system also features a “lazy registration” mechanism: new AIs do not need to register in advance; as long as someone pays it, the system will automatically create a profile. This lowers the entry barrier, allowing any AI program to start providing services and building a reputation immediately. All work records, evaluations, and ratings are permanently stored on the Solana blockchain, where anyone can view and verify them, but no one can tamper with them.

MoneyMQ: Turn the payment system into a configuration file

The last awarded project MoneyMQ is a developer tool, and its philosophy is that “payment systems should be as simple as writing configuration files.”

In Web2, if you want to add payment functionality to your application, you need to: register a payment service provider account, integrate their SDK, write code to handle various payment statuses, set up a testing environment, handle refunds and disputes… This process can take several weeks or even months. However, MoneyMQ simplifies all of this into “writing a few lines of YAML configuration files on your laptop.”

Imagine YAML as a commodity, or rather, as a set of game rules; it would probably look like this:

Product Name: Advanced API

Access Price: 0.1 USDC

Billing method: Based on the number of calls

You can write these rules locally, and MoneyMQ will automatically launch a complete payment environment, including product catalog, billing logic, test accounts, etc. You can simulate the entire payment process on your own computer: initiate payment requests, verify the x402 protocol, and check fund arrival. When the tests are fine, deploy to the production environment with one click, and all configurations will take effect automatically. MoneyMQ has built-in support for the x402 and MCP protocols. This means that AI programs can not only use your services but also understand your billing rules and even help you optimize pricing strategies. For example, the AI can analyze “how much the call volume would increase if the price is reduced from 0.1 USDC to 0.08 USDC,” and then suggest you adjust the price.

The “Embedded Yield” feature planned for the project launch is also very creative: the balance in your account will not remain idle, but will automatically participate in DeFi (Decentralized Finance) yield strategies. For example, if you earned 1000 USDC this month, this amount will automatically earn an annual yield of 4-5% before you decide to withdraw it. This is a considerable additional income for companies with large cash flows.

MoneyMQ has already provided a Homebrew installation package for macOS, allowing developers to install it with a single command.

written at the end

Of course, these projects are still in their early stages, but the possibilities they showcase are already quite exciting. For the average user, these technologies may still seem a bit distant. But just imagine: perhaps in the near future, your smart home system will purchase weather forecast services on its own to decide whether to water the plants, your dashcam will sell captured traffic information to mapping companies, and your health monitoring wristband will pay to use the latest AI diagnostic models… When AI is able to autonomously handle these micro-payments, our digital lives could become smarter and more convenient.

The organizers stated that the winners of the partner track will be announced next week.