MatrixOps

• 

  homepage (all simulations are just based on example data only, no real data)

• 

  Food Grocery Service Sections

• 

  Choose Food Category, enter your requirement by using chatbox, voicechat and image upload.

• 

  MatrixOps finds your best 3 options that aligns with your requirements and stated the reason of MatrixOps chosen this options

• 

  After selected one options, MatrixOps direct you to next page to confirm your options whether aligns with requirements.

• 

  After comfimed your options, MatrixOps will bring your to final confirmation page before directing you to the payment page.

• 

  After finished the payment by your own verification, you will be directed back to MatrixOps.

• 

  When choosing grocery category, above procedures are same applicable to grocery options.

• 

  When choosing service category, above procedures are same applicable to service options.

• 

  What if there are no options align with your requirements, MatrixOps will allow you to increase the budget to find the most relevant option.

• 

  MatrixOps will list the most suitable options that can fulfil your requirements.

• 

  Users can also press the try another scenario button to access What-if simulation features.

• 

  Users can change your budget so MatrixOps will simulate it to find you the new options

• 

  If you need to order food for your friends who have different requirements from you. You can choose decide multiple people at the bottom.

• 

  MatrixOps will help to find best compromise following each priority to avoid multiple ordering if possible.

Inspiration

I found that people spend a lot of time searching for food, either physically or online. Sometimes, even when they reach the restaurant or the confirmation page for an order, they suddenly cancel the order because there are no promotions, they are concerned about the price, or they forgot about their allergies or vegetarian preferences.

Some people who go to the gym need to control their sugar and protein intake strictly. Moreover, I live in Malaysia, and I have seen some of my friends having to access different websites and platforms to verify whether the food they order is halal or not. Even for me, I always look for the price before ordering food, but it takes me almost half an hour.

So, I came up with the idea: why not have a platform to help us search for the best options, spend only a few minutes, and at the same time fulfil most of our requirements?

After I heard about the OpenAI WebMCP Challenge, I tried to understand what a WebMCP tool is. It is an AI agent that can help humans make decisions using natural language. Also, AI's feature of finding exact information from different databases allows a platform to gather information easily and find what is really needed. So, with AI and WebMCP, here is why MatrixOps came out now!

What it does

So, basically, by using an AI agent and WebMCP tools, MatrixOps can truly help users find their preferred options by finding the top 3 options that fit their requirements for food, groceries, and services from a wide range of options. However, MatrixOps current website uses synthetic demonstration data and hasn't connected to real market data yet. There are a few features such as uploading images, requirement negotiation, what-if simulation, multi-person decision-making, and options trending which can help users to get their preferred options when they encounter issues while using the website. WebMCP allows AI agents to interact with websites through structured tools using natural language. AI agent and WebMCP tools can collaborate with users to make the decision but also leave the final decision to the user for human confirmation.

Hopefully, if MatrixOps can really be launched in the market, I hope users can purchase what they need in a very short time while fulfilling their requirements.

How we built it

So, after I came up with the idea, I shared my proposal with ChatGPT. It helped to sharpen and finalize the idea.

Next, ChatGPT provided me with the code so that I could start my vibe coding for MatrixOps in VS Code. I then implemented WebMCP tools for food, grocery, service searching, option comparison, and preparing an order. This let MatrixOps have a better performance in helping users in decision support. After spending plenty of time, I finally finished MatrixOps with the features, functions, and requirements that I had predicted before.

Next, I pushed all my files to GitHub as a repository and deployed the website using Vercel.

Challenges I ran into

I have faced many problems during the creation of my website, especially in coding. ChatGPT helps me a lot with coding. It helps me structure my code and find out every error that exists.

I have also learned that coding is a very complicated task, as it takes a lot of time to run simulations to ensure whether the code works properly and matches my expectations. Also, there are really a lot of errors that can be found during coding. And I also met issues in running WebMCP tools when try to deploy MatrixOps in Chrome and ChatGPT built-in browser. Luckily, I am able to solve it with my partners -- Chatgpt.

There were a few times when I felt burnt out because there were numerous errors. Hence, it motivates me to start learning programming languages to have a better coding experience.

Accomplishments that I proud of

So, after spending three days, I finally finished MatrixOps. I felt very satisfied to have created my own website, which I had never imagined before. Although throughout the journey there is stress from running out of time, I found that I actually quite enjoyed the moments and was fully engaged in brainstorming my ideas into reality.

This challenge has inspired me to continue exploring hackathons and building some products because this is the second time for me to take part in a hackathon.

What I learned

I have learned to use GitHub, VS Code, Vercel, and most importantly, ChatGPT. Usually, I use ChatGPT to ask for advice, do my research, and work on projects.

This is a very new experience for me to build a website with the aid of AI tools. I have also learned that coding is a very complicated task and that there can be numerous errors during development.

Hence, it motivates me to learn programming languages to have a better coding experience. I have also learned about the applications of WebMCP tools, I hope I can apply them in my undergraduate studies and building products.

What's next for MatrixOps

For future development, I hope to design it as a multilingual website. Actually, a small part of the website is currently available in multiple languages, but due to time constraints and the difficulty involved, I hope to continue developing it to make it more user-friendly.

Also, there are a few features that can be further developed, such as option ranking and a cross-category decision engine, which allows users to plan their activities by setting a budget. Features such as requirement negotiation which handles conflicts among multiple options and What-If simulation which allows users to adjust their budget when no options meet requirements, can also be further advanced.

In the future, I also hope to add a new function which is a scheduling feature. For example, if users already know what they want to eat every Monday or Friday, they can save their preferred options and schedule them in MatrixOps. With real data and fully implemented WebMCP-enabled services, the AI agent could directly connect to the ordering page, so users would not need to search and compare everything again. They would only need to review and confirm the payment.

Most importantly, the database of the website currently contains only example data. Hopefully, I can connect it with real data from the market if I get the opportunity.

Built With

• ai
• chatgpt
• github
• html
• javascript
• vercel
• vs
• webmcp