Horizon Trade
type a sentence -> trade the market
I wanted to share what we’ve been building over at Horizon Trade and get some honest feedback from the community.
For a long time, institutional-grade algorithmic trading was a luxury reserved almost exclusively for Wall Street firms and quantitative analysts with advanced computer science degrees. If you couldn't code custom infrastructure in Python or master proprietary languages like Pine Script (TradingView) or MQL5 (MetaTrader), you were essentially locked out of automated markets. You were stuck manual "screen-staring," battling emotional fatigue, FOMO, and panic-selling.
The barrier keeping everyday investors from competing on this level isn’t a lack of market insight; it’s a steep technical roadblock. We wanted to permanently tear down that wall.
What is Horizon Trade?
Horizon Trade is an advanced, no-code AI algorithmic trading platform designed to democratize sophisticated quantitative analysis.
Instead of spending weeks writing and debugging syntax, our underlying AI architecture acts as an intuitive translator for your financial ideas. You can build, validate, and deploy automated trading strategies using plain English.
For example, you can literally type a prompt like:
"Buy when the 15-minute MACD crosses above the signal line while the RSI is below 40, and set a 2% trailing stop loss."
The platform instantly converts that natural language thought into a structured, rule-based algorithmic trading system.
The Core Workflow We Built:
No-Code AI Generation: Write your strategy in everyday text; our LLM translation layer generates the exact logic parameters.
High-Fidelity Backtesting: Before risking capital, you can stress-test your ideas across multiple asset classes (stocks, crypto, forex) against standardized quantitative metrics like maximum drawdown and Sharpe ratio, yielding a proprietary "Horizon Score."
One-Click Execution: Run your strategies 24/7 on autopilot via secure, native broker integrations (Alpaca, Binance, Coinbase, etc.).
The Tech Challenge We’re Navigating:
Building an LLM infrastructure that doesn't just guess, but accurately translates human language into flawless, deterministic execution logic without syntax errors has been an incredible engineering hurdle. We’re constantly refining the data pipeline to ensure high-fidelity historical data matches live execution perfectly.
We are building Horizon Trade to shift the balance of power in the markets, giving individual traders the exact same tools that the world's largest funds enjoy.
We’re currently optimizing the user workflow and the UI. I’d love to hear your thoughts:
If you trade, what's the biggest friction point keeping you from automating your strategies?
For the devs here, what are your thoughts on using natural language processing to generate rule-based financial models?
Check us out at horizon.trade and let me know what you think!
2 Comments
I saw this pattern before with natural language to trading tools. the hard part is never strategy generation, it is live execution behaving the same as backtests.
Most teams underestimate slippage, edge decay, and how quickly paper performance stops matching reality.
Without strong real world track record signals it stays in the interesting but hard to trust category.
Reading this, I found myself less curious about the translation layer and more curious about the assumption underneath it.
The post makes a strong case that implementation is the barrier.
What I'd be hesitant to assume too quickly is whether implementation and adoption are constrained by the same thing.
Those can sometimes look surprisingly similar while leading products in very different directions.