The AI investing stack

Multi-agent LLM trading framework modelled on a trading-firm org chart

Open-weights reasoning model widely used as a low-cost brain for trading agents

LLM-driven watchlist analysis for A-share, HK, US, Japan, Korea and Taiwan stocks, pushed daily

A team of investor-persona agents that debate a stock and size a paper position

The most popular open-source crypto trading bot, with FreqAI for ML-driven strategies

Microsoft's AI-oriented quant investment platform, from data to backtest to execution

One API for 100+ crypto exchanges, used by Freqtrade and most crypto bots

HKU's open-source multi-agent research workspace: natural-language research, backtests, skills and MCP tools

Chinese-language fork of TradingAgents extended for A-shares and Hong Kong

Rust-native, event-driven trading platform with identical backtest and live code paths

The ubiquitous Python library for Yahoo Finance data, and the default feed in most agent demos

The long-standing Python backtesting and live-trading framework

HKU's agent-native trading platform: any agent registers by reading a skill file and trades on a shared venue

Institutional-grade event-driven backtesting and live-trading engine

Open-source financial LLMs fine-tuned with LoRA on news, filings and social data

Code for Stefan Jansen's book: the most complete open ML-for-trading curriculum

Open-source framework for market making and algorithmic trading on crypto exchanges

The reference deep-reinforcement-learning framework for trading and allocation

Microsoft's research-and-development agent, with a quant factor and model loop built on Qlib

Self-hosted AI trading terminal with agents that trade crypto, equities, FX and commodities

Open-source "AI trading OS" combining agent trading with a strategy workbench

Python wrapper for the standard technical-analysis indicator library

Gallery of deep-learning stock-forecasting notebooks from the pre-LLM era

Vectorised backtesting on NumPy and Numba for testing thousands of parameter sets at once

Small, readable Python backtesting library for single-asset strategies

Crypto research and trading framework focused on strategy development

Open-source AI agent platform for financial analysis from the AI4Finance Foundation

A six-month real-money micro-cap experiment run by ChatGPT, with the full forward-only record and an evaluation paper

TypeScript AI trading agent that covers equities, crypto, commodities and FX

Long-running curated list of LLMs, deep-learning strategies and tools in finance

Free-tier API for prices, fundamentals, FX and technical indicators

Portfolio optimisation library with a very wide set of risk measures

Collection of agent skills for financial analysis following the Agent Skills open standard

Claude Code skill workflows for individual investors: market review, risk, trade planning and journaling

Autonomous AI trading agent spanning prediction markets, crypto exchanges and more

Read SEC EDGAR filings in Python, including parsed XBRL financial statements

Portfolio optimisation built on the scikit-learn API, with proper cross-validation

Self-improving trading agents whose prompts are rewritten when they lose, from General Intelligence Capital

MCP server exposing the Financial Datasets stock-market API to AI assistants

The classic BERT model fine-tuned for financial sentiment classification

Hedge-fund-style agentic trading stack for crypto with a hard Risk Guard veto before any order

Python trading framework that backtests and places orders through a dozen brokers

Maintained fork of Quantopian's Zipline backtester

Dynamic datasets and market environments for FinRL research

"Claude Code for financial markets": agent teams that research, keep a thesis workspace and trade within set permissions

Global multi-asset brokerage with TWS, Client Portal and FIX APIs

API-first broker with commission-free stocks, options and crypto and an instant paper account

Decision audit trail and persistent memory layer for AI trading agents

Research-only multi-agent market research platform with scouts, reviewers and a separate execution boundary

Alpaca's official MCP server: place and manage orders from an AI assistant

LLM trading agent with layered memory and a tunable risk character, from the paper of the same name

Financial instruction-tuned LLM plus the FinBen evaluation benchmark

Curated list of LLM trading agents, MCP servers and agent skills, organised by building block

7B reasoning model trained for financial problem solving with SFT and RL

Academic platform for prototyping, backtesting and paper-trading LLM agents, built alongside a survey of the field

Open-source alternative to Yahoo Finance data, served from a Hugging Face dataset, with an MCP server and skills

Kraken's official AI-native CLI: full exchange API, built-in MCP server, live and paper trading in one binary

Minimal Python package for bulk-downloading SEC filings

Multi-agent trading system where agents compete internally before one view drives the decision

Toolkit for AI-driven strategies on Kalshi prediction markets

Benchmark of open-book financial questions answered from company filings

Code for the book on AI trading with Python, QuantConnect and AWS

Normalised tick-level market data with a pay-as-you-go model

Harness for creating and managing AI trading agents on top of Hummingbot

UIUC's live-market evaluation of LLM trading agents across US equities and Polymarket

MCP server for Twelve Data's real-time and historical market data

Paper artifacts and aggregate data from a continuous live record of LLM trading agents

Anthropic's frontier models, common in document-heavy financial analysis

The default LLM provider in most open-source trading agents

Google's multimodal models, used for research agents and large-context document work

The 50-billion-parameter proprietary model that started the financial-LLM race

Open data platform for analysts, quants and AI agents, with an MCP server built in

Classical and modern portfolio optimisation in Python

No-code platform that turns natural-language ideas into automated, rebalanced strategies

Official MCP server for Massive market data, formerly Polygon.io

Official REST and WebSocket client for Massive market data

Broad fundamentals and pricing API with an MCP server and generous coverage

Brokerage API with a sandbox and flat-fee options pricing aimed at developers