Today we are launching QuantForge, a self-hosted crypto trading platform built around a simple idea: your AI should be a teammate, not a widget.
Chat-First Architecture
Most trading platforms bolt AI onto existing interfaces as an afterthought. QuantForge inverts this. The AI chat is the primary interface. You tell it what you want in plain language, and it calls internal tools to make it happen.
Want to backtest a strategy? Ask. Want to deploy a bot? Ask. Want to check risk exposure across your portfolio? Ask. The AI has access to 15 internal tools spanning backtesting, bot management, risk monitoring, market data, and strategy analysis. No dashboards to navigate, no forms to fill.
40 Strategies, 6 Categories
QuantForge ships with 40 battle-tested strategies across classical TA, derivatives, macro, on-chain analytics, statistical methods, and ML/adaptive approaches. Each strategy has been validated through our 5-stage pipeline to filter out overfitting before deployment.
You do not need to understand every strategy. The tournament system screens them automatically, ranks by risk-adjusted returns, and surfaces the winners for your specific symbols and market conditions.
Self-Hosted by Design
Your API keys never leave your machine. Your trading data stays on your hardware. QuantForge runs as a single Docker container with SQLite for storage, FastAPI for the backend, and React for the dashboard. No cloud subscriptions, no vendor lock-in, no third-party risk.
The platform ships with pre-trained ML models, seeded historical data covering 25 symbols and 20 million candles, and pre-configured paper trading bots so you can see results within the first hour.
What is Next
We are focused on three areas: expanding the derivatives data pipeline with options volatility surface data, building a portfolio optimization layer that dynamically allocates capital across strategies, and shipping one-click deployment for common cloud providers. Follow the blog for updates.