Self-hosted · C++20 · MIT

The open-source options trading simulator

Practise prop-style evaluations on real option chains, from Cboe’s free delayed data or your own provider, with analytics you can check.

Try it in one command Read the source

Your API keys, your data and your trades stay on your machine.

The demo market’s five simulated days, then SPX and QQQ options trading a simulated day at 120 times real time. — Generated days to trade when the market is closed. A reversal, a trend, a chop, a selloff and an overnight session. Collapse the demo to hide the animation.

Start here

Try it in one command

With Docker installed, start Cboe’s free 15-minute delayed feed for SPX, SPY, QQQ, IWM and DIA. No key needed.

docker run --rm -p 127.0.0.1:8080:8080 -v openport:/var/lib/openport ghcr.io/38st/openport

Open the link it prints, http://localhost:8080/#token=…. The volume keeps your accounts and chart history between runs. When nothing is trading, choose the demo market’s simulated prices.

Use your provider, a release archive or a source build

The simulator

Trade like an evaluation

A profit target and a trailing drawdown floor decide pass or fail. Build multi-leg strategies, set brackets, and review each attempt in the journal. Record a session and trade it again in replay, or start a simulated day in the demo market.

Orders fill against displayed quotes. On the default delayed feed, a pass is practice, not proof: a real-time chart shows where the market went next.

Rules, fills and simulation limits

The analytics

Analytics you can audit

OpenPort computes its own implied volatility and Greeks, fits SVI surfaces, and shows gamma and vanna exposure. The forward and discount factor come from put-call parity; the provider’s IV stays alongside for comparison.

IV solve: About 0.4 µs 5.4 Newton iterations per option.
SPX analytics pass: About 40 ms 30,182 options, 63 expiries.

Timings are medians on an Apple M2 Max on 2026-09-24: the IV solve from openport_bench, and the SPX pass over 23 passes on live data during the session. SVI fits run separately, on demand in the API.

Against Cboe’s published IVs on 2026-09-24 during the session, median differences were 0.012 vol points for SPX, 0.030 for QQQ and 0.028 for SPY, across every expiry on out-of-the-money options within 10% of the forward.

Exposure assumes dealers are long calls and short puts: a modelling convention, not knowledge of anyone’s positions.

Methods, assumptions and measurements

On your machine

Your data stays yours

One C++20 binary runs the feed, analytics, simulator and web terminal. Self-host it with your own keys. Start with Cboe’s free delayed quotes, or connect Databento, Massive, ThetaData, or your Tradier or tastytrade account’s market data.

These adapters follow their documented APIs and are tested against sample responses, but have not yet been run live with a key or account. The broker adapters read market data only; they never send orders.

Write a provider adapter

Inside the terminal

From the chain to the whole book

Recorded views of the terminal, not a live feed. All trading and P&L are simulated.

Trade view in dark theme: SPX five-minute candles, option chain and simulated order ticket on Cboe’s delayed data. — **Trade.** The underlying’s chart, option chain and order ticket together. Cboe delayed data; simulated trading.

Volatility view in light theme with SPX smiles, term structure and parity forwards. — **Volatility.** SPX smiles, term structure and forwards, in the light theme.