QuantBase

QuantBase Dashboard
QuantBase Bot Creation Panel
QuantBase Logo: Garry the Duck

Inspiration

QuantBase is democratizing algorithmic trading. We believe that every person deserves access to wealth building tools that only major banks and hedge funds can have access to. Using QuantBase you can develop bots through natural language interactions with an LLM, no coding required. These bots can be displayed on the marketplace, where other users can rent them to use for live trading, earning the creators commission. Users can also discover, test or deploy bots from other creators in seconds.

What it does

Explore our marketplace of trading algorithms, see their historical performances, and implement them on our test blockchain with a single click.
Use our fine-tuned Claude LLM, adjusting for risk tolerance, leverage, and capital.
Upload your custom trading bot to our marketplace, allowing others to download and use it while you earn commission.

How we built it

We used Solana Faucet and Devnet to facilitate the trades. Under the hood, QuantBase supports various trading strategies, including Momentum and Mean Reversion Trading. Our fine tuned Claude LLM allows the users to modify algorithmic parameters, balancing between aggressive and conservative behaviors depending on the capital and leverage preferences. The frontend is built with Next.js and shadcn/ui for a modern and beginner-friendly interface. The entire project is hosted on Vercel.

Challenges we ran into

Our biggest challenge by far was to store, display and compute the sheer number of data that our program produces. QuantBase is a HFT (High Frequency Trading) platform that processes multiple real-time data streams such as our models, real time exchange trading and much more. The scale grows exponentially as we introduced more features like custom bot creation and time forecasts. To solve this issue, we built an extensive Python backend powered by MongoDB for data storage, retrieval, and high-speed computation.

Another one of the biggest challenges of the project has been to set up our own locally hosted crypto developer net. Since our trading models utilize HFT, we couldn't use the Solana developer net due to frequency limits. We had to configure a local validator and custom RPC environment to simulate real time trading.

What's next for QuantBase

We are committed to the idea of democratizing Quantitative Finance. Everyone should have access to the wealth building resources that the 1% has access to and we are motivated to build and develop QuantBase to be the app that facilitates that. Our next steps are to deploy QuantBase on the Solana mainnet and expand Claude-powered LLM features for smarter financial model building. We plan to add more trading algorithms and deepen AI-driven customization to make algorithmic trading accessible to anyone.