AI Trading

Kronos is an MIT-licensed, decoder-only foundation model pre-trained on K-line (OHLCV) data from 45+ global exchanges, accepted at AAAI 2026, that turns candlestick forecasting into a language-modeling problem you can run, finetune, and backtest yourself.

Vibe-Trading is an open-source research workspace from HKUDS (HKU Data Science) that turns natural-language finance questions into backtests, multi-agent analyst reports, and optional broker-authorized trading.

An LLM-driven multi-market stock analysis system with real-time news, decision dashboards, and automated notifications. 48K stars. Supports zero-cost scheduled runs.

AI-Trader is an agent-native trading platform from HKUDS that enables AI coding agents like Claude Code, Codex, Cursor, and OpenClaw to autonomously execute trades, manage portfolios, and optimize strategies.

Freqtrade (51,300 GitHub stars) is the open-source crypto trading bot written in Python. Backtest strategies, optimize with hyperopt, deploy to exchange APIs. Includes setup guide, strategy development, and real backtest benchmarks.

CloakBrowser (25,077 GitHub stars) is a stealth Chromium that passes every bot detection test. Drop-in Playwright replacement with source-level fingerprint patches. 30/30 tests passed. Includes setup tutorial, anti-detection breakdown, and benchmarks.

TradingAgents is an open-source LLM multi-agent framework (82,254 GitHub stars, Apache-2.0) that simulates a trading firm: analyst, researcher, trader, and risk agents debate a BUY/SELL/HOLD call. Built on LangGraph. Covers install, the agent pipeline, CLI + Python API, and an honest comparison with Qlib and single-agent bots.

Ran 7 quant experiments, found "textbook overfit" (Train PF 2.08 → OOS 0.94, ratio 2.21). Then discovered the diagnosis itself was wrong — silent schema field mismatch made the optimizer run with default 10x leverage instead of the evolved 2x. The corrected version is healthy (ratio 1.01). The meta-lesson is uglier than the original.

Hands-on review of moss-trade-bot-skills v1.0.26: a natural-language quant agent builder for Hyperliquid perps. Industrial-grade backtest engine with Decimal precision and depth-book modeling — but a Sharpe annualization bug and a textbook OVERFIT trap once you enable evolution. Setup walkthrough, OOS validation results, and the honest take on whether evolved params survive out-of-sample.

After 50+ live trades from optimizer outputs, we cataloged 5 distinct overfit patterns: walk-forward divergence, regime-flip, parameter-cliff, indicator-stacking, and survivorship. Each with reproducible synthetic example + the detection signal.

Complete guide to Moralis Web3 Data API in 2026. Learn how to fetch real-time blockchain data, NFT metadata, token prices, and wallet balances across 10+ chains with JavaScript, Python, and Unity SDKs.

Master CCXT, the #1 open-source crypto trading library. Connect to 100+ exchanges with one unified API. Build Python trading bots with real-time WebSocket data, built-in rate limiting, and backtesting support.

Complete guide to the Alpaca Trading API for commission-free algorithmic trading. Learn setup, order placement, WebSocket streaming, fractional shares, and paper trading with Python code examples.

Master VectorBT for quantitative backtesting in Python. Build, test, and optimize trading strategies with vectorized Numba-accelerated simulations. Complete 2026 guide with code examples.

Master TensorTrade for RL-based algorithmic trading. Build custom Gym environments, integrate Stable Baselines3, and deploy production-ready portfolio management strategies with real benchmarks.

Complete guide to TA-Lib Python wrapper with 200+ technical indicators. Install, benchmark, and deploy SMA, EMA, RSI, MACD, Bollinger Bands for algorithmic trading in 2026.

Build production-ready PancakeSwap trading bots on Binance Smart Chain. Web3.py integration, automated strategies, liquidity pool monitoring, MEV protection, and Python bot framework — with real 2026 benchmarks.

Complete 2026 guide to Lean, the algorithmic trading engine behind QuantConnect. Multi-asset backtesting, live trading, C# & Python APIs, and production deployment walkthrough.

A production-ready guide to Jesse AI trading framework — install, backtest with 30+ indicators, build custom strategies, and deploy live crypto trading bots in Python.

A hands-on guide to deploying Hummingbot v2, the open-source crypto trading bot with 50+ exchange connectors. Covers Docker setup, custom strategies, backtesting, DEX gateway, and production hardening.

A hands-on guide to deploying Freqtrade with FreqAI, the open-source Python crypto trading bot with ML integration. Covers Docker setup, hyperparameter optimization, backtesting, Telegram integration, and production deployment.

Full guide to Backtrader event-driven backtesting engine. Build, test, and optimize trading strategies in Python. Integrations, benchmarks, and live trading deployment 2026.

Minara is the AI-native trading platform built on Hyperliquid that lets you ask questions, get real-time market analysis, and execute crypto / stocks / commodities trades in a single chat interface. Hands-on review: setup walkthrough, five real use cases, honest pricing breakdown, and how Spark token rebates stack with the 10% referral commission.

Hyperliquid is the fastest fully on-chain perpetual DEX with 100ms latency, native margin trading, and a powerful API for algorithmic and automated trading strategies.

CoW Protocol is a DEX aggregator that saves traders over $100M in MEV losses through batch auctions and solver competition. No sandwich attacks, no frontrunning.