HeyTraders
Personal AI quant research & backtest with natural language
148 followers
Personal AI quant research & backtest with natural language
148 followers
HeyTraders is a Personal AI Quant for crypto markets that translates natural language into verifiable trading strategies. Stop trading based on gut feelings and complex charts. Just talk to our agent to research trends, verify ideas with instant backtesting, and get data-driven confidence.
HeyTraders
Hey-Hunters! 👋 I'm Minseom, co-founder of HeyTraders.
While Soonbeom brings the quant expertise, I've been focused on making sure your data stays yours.
Here's our promise:
Your trading strategies are fully encrypted and stored securely.
We only use system data for stability improvements — never for training, never sold to third parties.
Your alpha stays your alpha.
We're just getting started, and your feedback means everything to us.
Try it out and let us know what you think — the good, the bad, all of it! 🙏
Wishing you all a great day ahead! 🌟
Interesting idea, would love to know if there will be a market for strategies later on, or even combo things together for better risk control
HeyTraders
@david_chen37 Thanks Siyuan, great ideas.
Marketplace: Yes, we envision an ecosystem where users can monetize proven alphas via a Marketplace and copy trading.
Combos: This is essential for risk management. We are planning a "Portfolio Backtest" feature to analyze correlations between strategies and see how they hedge each other.
Product Hunt
HeyTraders
@curiouskitty This touches the core challenge of any quantitative analysis.
You are absolutely right. Many traders get tricked by perfect backtests. Here is how we handle those specific assumptions:
1. Handling the Assumptions
Fees & Slippage: We apply fees and slippages default to 5bps (0.05%) to be conservative, but this is fully configurable. We also plan to integrate Orderbook depth replay later for realistic impact simulation.
Execution (Bar-close): If a signal is generated based on bar-close data, we strictly execute at the Next Open. We never assume you can fill at the closing price of the signal candle.
Lookahead/Repainting: Our engine strictly enforces time-series separation. The strategy logic only sees data up to time T, making lookahead or repainting impossible by design.
2. How to Audit (Why you can trust the curve)
You can trust the backtest by trade history. On backtest detail page, you can click the chart to inspect every single trade entry/exit on the candle. You can visually verify exactly where the bot bought and sold relative to the price action.
---
And more importantly, there are other major reasons why traders distrust AI Trading and backtesting. Here is how we address those specific fears:
Deterministic Protocol: LLMs are probabilistic, but trading must be precise. Our AI acts as a translator, not a creator. It converts your natural language logic into deterministic execution logic.
Trust Pipeline: We know backtesting is just a simulation and often fails in live markets. That is why we structured our flow as a verification pipeline: Backtest (Past check) → Live Signals (Real-time verification) → Automated Execution. We encourage you to verify the logic with Live Signals first. Only when you trust the real-time performance, should you enable automated execution.
Would love to hear your feedback on this approach!
The tool aims to deliver "data-driven confidence", which is far more reassuring than trading based on intuition. But I’m curious: beyond backtesting results, can it provide risk assessment metrics such as maximum drawdown or Sharpe ratio to help me evaluate a strategy in a more comprehensive way?
HeyTraders