AlphaBot

GIF
Architecture Diagram
PROFIT

Inspiration

The inspiration behind AlphaBot stemmed from the desire to capitalize on the fast-paced and often volatile Solana memecoin ecosystem. The goal was to create a trading bot capable of capturing market inefficiencies almost instantaneously, ensuring traders can seize opportunities before others and thus earn a GUARANTEED profit. To enhance the user experience, we also aimed to provide an intuitive interface through a Telegram bot.

What it does

AlphaBot is a high-performance crypto trading bot built with Node.js, designed to exploit inefficiencies in the Solana memecoin market. It listens for on-chain transactions in real time using Quicknode’s WebSocket, and buys memecoins within a second of detecting a relevant transfer. The bot also automates the selling process, making it a hands-free solution for traders looking to stay ahead in a fast-moving market. Additionally, we implemented a Telegram bot interface to control AlphaBot and view historical profits, leveraging AWS serverless technologies and containerized deployments.

How I built it

For the mid-frequency algorithm trading, we built AlphaBot using Node.js, connecting to a Quicknode RPC WebSocket endpoint to stream real-time on-chain transaction data. The bot subscribes to specific events and quickly processes incoming data, allowing it to respond with buy transactions in under one second. A crucial part of this development was navigating Quicknode's documentation to understand how to subscribe to data streams and parse transaction formats efficiently.

For the Telegram interface, we utilized a combination of AWS serverless and containerized technologies, along with Terraform infrastructure as code, GitLab CI/CD pipelines, and automated testing to produce a robust development workflow. To handle profit details, we implemented an S3 event triggered Lambda function that uses AWS Textract, an AI service to convert images to text, as the third-party trading platform only provides profit details in image format.

Challenges I ran into

We faced significant challenges while debugging the CI/CD pipeline, which frequently failed for unknown reasons. Although some team members had experience with CI/CD in the workplace, this is our first experience developing a full-fledged pipeline within the time constraints of a hackathon. Additionally, modifying the architecture to accommodate the third-party trading platform's requirements introduced complexity, especially in ensuring compatibility with their image-based data format.

Accomplishments that I'm proud of

Though AlphaBot has been operational for just a few hours, it's already racked up impressive profits, showcasing its lightning-fast ability to seize opportunities and dominate the memecoin market!

As perfectionists, we also challenged ourselves to implement a full-fledged CI/CD pipeline and deploy everything to AWS using Terraform. While this was definitely overengineering, especially in a hackathon setting, the experience of balancing speed and reliability proved to be incredibly valuable.

What I learned

First hands-on experience connecting to an RPC endpoint and working with a WebSocket. Navigating Quicknode’s documentation was crucial to understanding how to subscribe to real-time data streams and parse the incoming on-chain transaction data. This experience provided valuable insights into the intricacies of working with Solana’s transaction formats, optimizing the data flow, and ensuring lightning-fast responses to market movements.

Gained a deeper understanding of the AWS ecosystem by leveraging a combination of serverless, containerized, and event-driven architecture.

What's next for AlphaBot

The next step for AlphaBot is to focus on optimization. We plan to integrate multiple RPC node providers to ensure the fastest connection is always used, minimizing latency and improving overall performance. This will help AlphaBot stay competitive and reliable in the ever-changing memecoin market. We also plan to further improve the Telegram interface and continue fine-tuning the Lambda function for better image processing and text extraction.