Three steps, in the order you use QuantGPT: backtest and arm, connect and deploy, futuretest and simulate. Members learn them in QuantGPT University, where every chapter is a lesson, one button into the product and a three-question check. Everything on this page stays open to anyone: the lessons, the ramp and the guides.
The chapters in the order of the three steps. The QuantGPT chapters open on every plan, the QuantGPT Pro chapters on QuantGPT Pro. Your progress is saved on your account.
QuantGPT or QuantGPT Pro, 3 days free: card up front, cancel any time. University opens with your account.
5 backtests and 5 AI calls for the life of the account, no card. Pick a plan when you are ready.
Run a backtest, read it without fooling yourself, and arm your AI with our data. Start here, no account needed.
Arm your AI, become a quant, run the backtester. Three short playable lessons on real market data, no sign-in needed. About eight minutes end to end.
Ten minutes, one real strategy, real S&P 500 data since 2000. What a quant is, why software beats emotions, and your first strategy, animated end to end.
Run a study. One click on a Strategy Library preset, about one second of compute over decades of market history. Your first backtest is on the tape.
Every test ends one of three ways. DATA VALIDATED beat the benchmark on growth and on risk, and held on the years it never saw. TRADE-OFF beat it on one measure. NO EDGE found none. The full findings system.
Change one number. Run it again. Watch what moves.
Save your studies. They follow you across devices on your profile.
Ask your connected AI to rerun your saved study with a different start year and report what changed. It can search the library, run studies, change rules and report findings mid-conversation.
Your own code at your own broker, on paper first. Coming to QuantGPT Pro: the runner will host your own code and send its orders to your own Alpaca account, under your keys. We never decide a trade. It opens after review.
Download your criteria or hand them to your AI, and place any orders yourself at your own broker.
QuantGPT Pro. Issue a key, make one call from your own AI or code, and see the limits your plan carries.
QuantGPT Pro. What the runner will do and what it never does: your own code, your own Alpaca account, your keys, paper first.
Read what markets did after past events, run your book through them, and write rules that watch the feed for you. The lessons play without an account.
One war onset against every one on record: the S&P 500 after it, the sign test, and the ripple from its country to the rest of the world.
A rule, its backtest from the next close, the guard on your book, and the signal your AI reads. You decide what happens next.
In University, QuantGPT Pro. Count the real bets in a book, see what correlation does to them, and read your own.
One copyable block. Paste it into any Claude or ChatGPT chat and it walks you through QuantGPT. No MCP needed.
Once armed, your AI can call the guide tool itself and teach you with live runs instead of words.
What DATA VALIDATED requires, the walk-forward split, and the methodology stated in full.
The papers and open data behind the library, from one backtest to a research process. 91 sources, 23 re-run in the terminal.
Claude Code, Claude Desktop, and claude.ai, step by step.
The connector setup, step by step, plus the paste-pack fallback.
The Gemini app by URL, or Gemini CLI with a key, step by step.
One entry in mcp.json with your key, then a first backtest from Cursor's agent.
Paste any return series: every start date, every calendar year, the drawdown, then your own fee, taxes and slippage. Free, in your browser.
175 documented strategies from research papers, famous investors, and their letters.