Trade with code |
Your one-person Wall Street.
An open-source AI agent that gives you your own research desk, quant team, trading floor, and risk management — all running on your laptop.
Inspired by Cursor, Claude Code, and Manus.
Quick Start
Clone and build
Node.js 22+, pnpm 10+, and Claude Code CLI installed. That's all you need.
git clone https://github.com/TraderAlice/OpenAlice.git && cd OpenAlice
pnpm install && pnpm buildRun
Start the server and open localhost:3002. No API keys or config needed — the default setup uses Claude Code as the AI backend with your existing login.
pnpm devChat
Tell Alice what you want — analyze a stock, check your portfolio, place a trade. She reads markets, reasons through her strategy, and executes. Everything is file-driven: Markdown for persona, JSON for config, JSONL for conversations.
How Different Alice Is
Neither a quant bot nor a trading assistant — here's what makes Alice different.
Explainability
Alice doesn't optimize for "completing trades" — she optimizes for "explainable good trades." Before every action, she thinks, explores, tests, and calculates. The result is a decision you can actually read and understand. This dramatically reduces hallucination in her operations and means AI trading is no longer a black box.
Reproducibility
Through a complete chain of reasoning, Alice arrives at specific actions each round. Because the reasoning is thorough enough, running Alice multiple times — or having a different AI analyze her decisions — produces near-identical conclusions. Her behavior is reproducible, and for trading, that means stable.
Extensibility
Alice can use any frontier model, expand to new asset classes, and upgrade her internal tools at will. The reason is simple: she's an agent, not a workflow. We didn't pre-build any pipeline — just like the projects we pay tribute to.
How It Works
| Step | Cursor | TraderAlice |
|---|---|---|
| Understand | Read codebase | Read market data |
| Plan | Design solution | Analyze indicators |
| Test | Run tests | Backtest strategy |
| Execute | Push to git | Submit orders |
OpenAlice is file-driven — Markdown defines persona and tasks, JSON defines config, JSONL stores conversations. The same read/write primitives that power vibe coding transfer directly to vibe trading. No database, no containers, just files.
Alice is reasoning-driven. She doesn't follow a fixed pipeline — she thinks through each decision using real market data, news, and technical indicators. You describe your strategy in natural language, and Alice figures out the rest: which indicators to check, when to act, and how to manage risk.
The architecture is extension-based — each capability (crypto trading, securities, market data, news, brain) is a self-contained tool package. Add new asset classes, swap AI providers, or write your own extensions. Just like the projects we pay tribute to.
Agent Loop
Every round follows the same lifecycle — whether triggered by you or by Alice herself.
Research
- getFrontalLobe()
- getEmotion()
- globNews() / grepNews()
- equityGetProfile()
Analyze
- think(observations)
- plan(options, decision)
- calculateIndicator()
- secGetQuote()
Execute
- secPlaceOrder() stage
- secWalletCommit() commit
- secWalletPush() push
Monitor
- secWalletSync()
- secGetPortfolio()
- updateFrontalLobe()
- updateEmotion()
News Collection
RSS feeds run in the background, continuously ingesting articles into a searchable archive. Alice queries this archive during research with globNews() and grepNews().
Event-Driven Loop
Cron scheduler fires jobs into an append-only EventLog. Heartbeat subscribes and re-enters the agent loop autonomously — Alice monitors positions 24/7 and only speaks up when it matters.
Why Alice?
We wanted a name that anyone — from any country, any background — could remember. Alice's Adventures in Wonderland is a story known worldwide, and we wanted to build an agent named Alice to be your companion in Vibe Trading.
Actually, the name "TraderAlice" was also inspired by Trader Joe's, though we have nothing to do with grocery stores.
Trading is a rabbit hole. Through trading, people learn how companies operate, how industries work, how nations conduct diplomacy, how politics and elections take shape. We hope every user can be like Alice — leap into the rabbit hole and begin their own journey.
Our Vision
- →Most trades will be executed by Alice
- →Human traders can finally get a good night's sleep
- →Describe your strategy in plain English, let Alice handle the math
- →World peace
Ready to try?
Four commands to your own AI trading desk. Open source, MIT licensed, runs on your laptop.