Intelligent Contract Testing: Explore the Future of Smart Contracts

Matrix AI Network
4 min readFeb 14, 2024

Dear Matrixians and Blockchain Developers, Follow the introduction about Intelligent Contract, DOCS LINK our platform, that seamlessly translates natural language into Solidity code, is ready for testing. What can be a better CNY present than this? With three unique models available, we offer an unparalleled opportunity to engage with smart contract developers in ways never before possible.

Understanding Our Models:

Model 1: Vector Knowledge Base Approach (

  • Core Idea: This model organizes mainstream contracts into modules, subsequently arranging these by function into a vector knowledge base. This innovative approach allows for training based on a structured understanding of contracts’ functionalities.
  • Unique Advantage: Enhanced precision in contract generation, thanks to the organized knowledge base, allowing for targeted training and output.

Model 2: Open-Source Corpus Approach (

  • Core Idea: This model utilizes existing open-source smart contracts as the training corpus without extensive labeling of the data. This method mirrors the training of mainstream large language models, leveraging the breadth of available contracts.
  • Unique Advantage: Broad understanding of contract variations, providing a wide-ranging foundation that supports diverse contract generation.

Model 3: Hybrid Approach (

  • Core Idea: A blend of the first two models, this approach combines the structured learning from the vector knowledge base with the extensive corpus of open-source contracts. This hybrid model aims to balance the depth of targeted knowledge with the breadth of general contract understanding.
  • Unique Advantage: The best of both worlds, offering both precision and variety in contract generation, making it highly adaptable to complex and nuanced contract requirements.

Your Guide to Testing (Important, imagine you are using ChatGPT):

  • Be Detailed Provide as much important detail and context as possible. Essentially, make your descriptions clear and avoid being too vague. For example, instead of saying “write an ERC20 protocol,” say “write a protocol that implements the ERC20 interface, with each method fully developed. Finally, comment on the specific function of each method.
  • Use Separators to Clearly Indicate Different Parts of the InputSeparators such as triple quotes, XML tags, and section headings can help distinguish different sections of text that should be treated differently. This can help the model better understand the content. For example, add a security check feature within code separated by triple quotes, “””【code】”””
  • Specify the Steps Required to Complete the Task If a task can be broken down, it’s best to specify it as a series of steps. Clearly writing out these steps can make it easier for the model to implement them. For example, use the following step-by-step instructions in response to user input. Step 1 > Generate ERC20 protocol code. Step 2 > Add a feature to the code from Step 1 to include a 10% transaction tax rate.
  • Provide Examples This is the classic few-shot prompt; first, give the model examples to follow. For example, generate ERC20 code in the style of code separated by triple quotes: “””【code】”””
  • Break Down Complex Tasks into Simpler SubtasksBreaking down complex tasks into simpler subtasks can lead to better performance from the model. You can split the requirements into multiple single tasks, incorporating the results of each task into the background knowledge for the next task. For example:

Generate example code for the ERC20 protocol.

Based on the code separated by triple quotes, add logic for security checks: “””【code】”””

Why Does This Matter?

These models represent the cutting edge in smart contract development, each offering distinct advantages that cater to different needs and preferences in the blockchain community. By participating in the beta testing, you have the unique opportunity to explore these models in depth, providing valuable feedback that will shape the future of this platform. We encourage you to dive into each model, understand their unique characteristics, and see how they align with your specific needs and ideas for smart contract development. Your insights will be instrumental in refining these models for the broader blockchain community.

How to Get Started?

Choose any of the addresses (,, to begin your exploration. Approach each model with an open mind and consider how the underlying approach can influence the smart contracts you aim to develop. Share your findings, challenges, and suggestions with us, and let’s collaboratively push the boundaries of what’s possible in smart contract technology.

Expectation Management

  1. It is still in testing stage, the User Interface is not final, nor is the user experience. It is mainly for functions.
  2. Bugs or imperfections are expected. If you encounter of such, please jot down details with screen cap and send to
  3. It will take a little time loading the first time.

Join us on this transformative journey, and let’s unlock new possibilities together. Any partnership, please send an email to

The Matrix AI Network was founded in 2017. In 2023, we enter Matrix 3.0 blending neuroscience with our previous work to realize the vision of the Matrix films.


Website | GitHub | Twitter | YouTube

Telegram (Official) | Telegram (Masternodes)

Owen Tao (CEO) | Steve Deng (Chief AI Scientist) | Eric Choy (CMTO)



Matrix AI Network

The Matrix AI Network was founded in 2017. In 2023, we enter Matrix 3.0 blending neuroscience with our previous work to realize the vision of the Matrix films.