
Based Agent
A telegram bot that levergaes AI Agents to never miss an alpha, be it from a telegram group by using sentimental analysis or tracking a whale making returns while you are asleep.
About
The problem it solves Let's be honest, missing a juicy 100x moonshot hurts more than a breakup, so we being degens decided to build Based Agent telegram bot, that uses power of AI Agents, Sentimental Analysis, scoring metrics and automated algos to ensure that a degen earns even while being asleep. We built this for a TRUE DEGENS, where they can buy memecoins on wow.xyz, even before the bonding curve is completed / liquidity is locked Our project can be dissected into two features:- 1. Alpha Group Sentimental Analysis We all have those groups where the calls are good enough to bet on, but as a human we can't be perpetually available for trading unlike markets which are there 24/7, 365 days unlike tradfi markets. A good alpha does not stay an alpha, so apeing it early is the key to success. Hence, once the user subscribes to a particular group, our bot will ensure that every time an alpha is being shilled, identifying the nuances behind the texts and the AI agent making a calculated decision whether it would be a good buy or not, is the key to always being in toes to toes with those moonshot calls. 2. Whale sniping There are influential people who's wallets have been doxed and there are some whales who have proved to be profitable with certain low cap token trades, our AI agent can take enough calculated decision to ensure the user gets alphas from whales before anyone else does. Additional Features We are going to add support for more memecoin platforms like virtuals and pump.fun (cross chain basically), as we aim to finish this end-to-end, conviction being gambling / prediction are clear PMF in crypto industry from our POV.
How it's made
Challenges we ran into Our biggest challenge was very very different tbh, being on a public wifi, telegram banned our bot token on our devices for a brief till an hour ago of writing this, which is basically end of the hackathon, it took us a long time to figure out that the problem was not in our code but being on a public wifi, leaving us scratching our heads for 2 hours. A few challenges which I feel we faced were 1. Indexing historical data of wow.xyz on base as we needed a healthy pool of meme coins. 2.We tried an agent via langchain, it did not turned out well till we found a better way to optimize temperature on the base agent sdk itself