Skip to content

About

Free & open-source trading terminal: volatility cones, bands & options levels on live charts — MetaTrader 5 & Rithmic feeds, built-in simulated market, English & Arabic guide.

Topics

algorithmic-tradingday-tradingelectronfastapimetatrader5pythonrithmictradingtradovatevolatility

Resources

Readme
MIT license

Contributing

Contributing

Security policy

Security policy
Activity

Stars

31 stars

Watchers

3 watching

Forks

15 forks
Report repository
BranchesTags

Repository files navigation

Quantum Terminal logo

Quantum Terminal

A free, open-source trading terminal that draws probability models —
volatility cones, bands and options levels — on top of live charts.

English · العربية

Release CI License: MIT Python 3.10+ React 18 Platforms

Quantum Terminal on the built-in simulated market: an XAUUSD hourly chart with this week's and last week's volatility cones (GBM, MJD and Bates), a paper position with its stop-loss and take-profit lines, and the open-positions table

Running on the built-in simulated market — the weekly volatility cones are computed by the terminal itself. No broker account needed.

بالعربية: ‏Quantum Terminal منصة تداول مجانية ومفتوحة المصدر، ترسم فوق الشارت الحي المدى الذي يُرجَّح إحصائياً أن يتحرك فيه السعر، وتعمل على Windows وmacOS وLinux مع سوق محاكى مدمج لا يحتاج أي وسيط. اقرأ الشرح الكامل بالعربية، مع المتطلبات وطريقة التشغيل خطوة بخطوة ↓

Why Quantum Terminal?

Most retail platforms are built around lagging indicators such as RSI or MACD. Quantum Terminal shows the range price is statistically expected to travel instead: forecast cones from GBM, Merton jump-diffusion (MJD) and Bates models, volatility bands and options-market levels (gamma exposure, call/put walls, max pain), drawn right on the live candles.

It is a fast, local display terminal: heavy calculations run in the background, so the app itself stays light — live charts, order entry and position management, all on your own computer.

Quick start

If you already have Python 3.10+, four commands are enough. The simulated market works on Windows, macOS and Linux:

git clone https://github.com/wJubran/QuantumTerminal.git
cd QuantumTerminal/backend
pip install -r requirements.txt
python launcher.py --sim

The terminal opens in your browser at http://127.0.0.1:8502. There is no sign-up, no licence key and no paywall. For live prices without a broker, start with python launcher.py --provider yahoo (forex, metals, energy, grains, indices, crypto) or --provider binance (crypto); see Data sources. New to Python, or want MetaTrader 5? Follow the step-by-step guide below.

Requirements

What you need Notes
Operating system Windows 10 or 11, macOS, or Linux (64-bit) MetaTrader 5 needs Windows; the simulated market runs everywhere
Python 3.10 or newer, 64-bit Tested with 3.10, 3.12 and 3.13. On Windows, tick Add python.exe to PATH in the installer
Memory 4 GB of RAM The backend itself uses about 150 MB
Disk space About 1 GB free The Python packages take about 500 MB
Browser A current Chrome, Edge, Firefox or Safari Or use the desktop app instead
Internet To install; afterwards for the live data sources, the macro data (FRED, CFTC, Yahoo Finance) and the update check The simulated market and CSV files work offline
MetaTrader 5 (optional) The MT5 desktop terminal from your broker, logged in to an account (a demo account is fine) Windows only. One way to get real prices and place live orders; see Data sources for the others
Node.js (optional) 20 or newer Only for the desktop app or to rebuild the UI
Git (optional) Any version Or download the code as a ZIP file

Installation, step by step

1. Install Python

  • Windows: download the 64-bit installer from python.org and tick Add python.exe to PATH on the first screen. Then open a new Command Prompt and check that python --version prints 3.10 or newer.
  • macOS: use the installer from python.org or brew install python@3.12. On macOS and Linux, type python3 wherever this guide says python.
  • Linux: install your distribution's packages, e.g. sudo apt install python3 python3-venv python3-pip on Debian or Ubuntu.

2. Get the code

git clone https://github.com/wJubran/QuantumTerminal.git
cd QuantumTerminal

Without Git: on the repository page click Code → Download ZIP, extract it, and open a terminal in the extracted folder (it may be called QuantumTerminal-main).

3. Create a virtual environment (recommended)

A virtual environment keeps the terminal's packages separate from the rest of your system. Create it in the project folder, where the desktop app also looks for it:

:: Windows (Command Prompt)
python -m venv .venv
.venv\Scripts\activate
# macOS / Linux
python3 -m venv .venv
source .venv/bin/activate

Your prompt now starts with (.venv). Activate it again in every new terminal window before you start the terminal. If PowerShell refuses to run the activation script, run Set-ExecutionPolicy -Scope CurrentUser RemoteSigned once, or use Command Prompt.

4. Install the dependencies

cd backend
pip install -r requirements.txt

The first install takes a few minutes. The MetaTrader5 package is only installed on Windows, because MetaQuotes publishes it for Windows only.

5. Start the terminal

On the simulated market (any system, no broker account):

python launcher.py --sim

On live prices, without a broker (any system, internet only):

python launcher.py --provider yahoo      # forex, metals, energy, grains, indices, crypto
python launcher.py --provider binance    # crypto, in real time

Orders then go to a paper account that fills them at those live prices. OANDA, Alpaca and CSV files are set up in Settings ⚙ → DATA SOURCES; see Data sources.

With MetaTrader 5 (Windows):

  1. Install MetaTrader 5 from your broker and log in to your account (a demo account is fine).

  2. To place orders from Quantum Terminal, turn on Algo Trading in MetaTrader 5 (the toolbar button, or Tools → Options → Expert Advisors → Allow algorithmic trading).

  3. Leave MetaTrader 5 running and start:

    python launcher.py

  4. If the terminal does not find MetaTrader 5 on its own, set it under Settings ⚙ → PROVIDERS: TERMINAL PATH (the path to terminal64.exe) or LOGIN CREDENTIALS (server, login and password).

The launcher prints the server address and where it keeps its config, data (AppData:) and log file, then opens http://127.0.0.1:8502 in your browser. If the browser does not open, open that address yourself. Keep the launcher window open while you use the terminal.

6. First steps in the terminal

  • Cones: on a chart, open the ▾ next to CONES and pick DAILY, WEEKLY, MONTHLY or QUARTERLY (and BATES for the Bates model). The first time, they are ready about 30 seconds after start-up.
  • Trading: orders are blocked until you switch on Settings ⚙ → TRADING → LIVE TRADING. On the simulator, and on the sources that only deliver prices, that switch only unlocks the paper account (100,000 USD, 1:100 leverage).
  • Data source: Settings ⚙ → DATA SOURCES switches between MetaTrader 5, Yahoo Finance, Binance, OANDA, Alpaca, CSV files and the simulator while the terminal runs. The choice is remembered.
  • Learning Center: the LEARN tab explains every page and model, in English and Arabic.

7. Stop, restart and update

  • Stop: press Ctrl+C in the launcher window, or close it.
  • Restart: open a terminal in the project folder, activate the virtual environment (step 3), then run cd backend and python launcher.py (add --sim for the simulated market).
  • Update: in the project folder run git pull, then pip install -r backend/requirements.txt with the virtual environment active. Your settings and the paper account are kept: they live in the AppData: folder, not in the project folder.
  • Uninstall: delete the project folder and the AppData: folder printed at start-up.

Desktop app (optional)

With Node.js 20 or newer installed:

cd electron_shell
npm install
npm start              # or: npm run start:sim

The app starts the backend for you, using the project's .venv (or venv) when it exists and the Python on your PATH otherwise, and shows the terminal in its own window. It keeps its settings in a separate folder (QuantumTerminal-v2) from the browser version. Package it with Electron Forge or electron-builder to distribute it.

Troubleshooting

Problem What to do
'python' is not recognized (Windows) Run the Python installer again and tick Add python.exe to PATH, or type py instead of python
error while attempting to bind on address Port 8502 is in use, usually by another copy of the terminal. Close it, or start with --port 8600
No MetaTrader5 package on macOS or Linux Expected: it exists for Windows only. Use --sim
MT5 initialize() failed in the log Check that MetaTrader 5 is open and logged in, and set TERMINAL PATH under Settings ⚙ → PROVIDERS
MetaTrader 5 rejects orders because algo trading is disabled Turn on Algo Trading in MetaTrader 5
The browser did not open, or the page stays empty Open http://127.0.0.1:8502 yourself, and look at the log file the launcher printed
No cones on the chart Wait about 30 seconds after the first start, then turn them on from the ▾ next to CONES
A ticker stays empty on Yahoo Finance, OANDA, Binance or Alpaca The source knows it under another name: set its broker symbol under Settings ⚙ → UNIVERSE (ticker names)
TEST in DATA SOURCES says 401 or 403 The token or API key is wrong or has expired, or a practice token is being used on live (or the other way round)
TEST cannot reach the source Allow its address in your firewall or proxy: query1.finance.yahoo.com, api.binance.com, api-fxpractice.oanda.com (api-fxtrade.oanda.com for live) or data.alpaca.markets

Features

Works out of the box

Area What you get
Volatility cones DAILY, WEEKLY, MONTHLY and QUARTERLY cones from GBM, MJD and Bates models, computed by the terminal from the feed's own daily bars (how)
Probability bands Rolling GBM, MJD and Bates bands that restart at every trading day's open, with their width taken from the last 60 trading days and a 20-day forward fan (how)
Replay Plays back past weekly, monthly and quarterly cones and the daily bands against price, bar by bar
Pulse Room Scanner that ranks your watchlist by where price sits inside its cones and bands
Regimes Each instrument's volatility regime (calm → stressed) from a walk-forward HMM, a 15-minute volatility alarm, and the macro growth/inflation quadrant from FRED data (how)
Fundamental Macro dashboard from public data: the regime quadrant, Fed net liquidity, the yield curve, the VIX, CFTC futures positioning and a score per instrument (how)
Money Flow Rotation between asset classes, US sectors and countries, from ETF prices and volumes (how)
Live charts TradingView Lightweight Charts, M1 to D1, several charts in tabs, drawing tools, chart templates, crosshair price/time readout, world clock with trading sessions
Order entry Market, limit and stop orders with stop loss / take profit, and a live risk / reward preview in money and in % of the balance
Positions Open positions with risk, reward and P&L; drag SL/TP lines on the chart, or edit and close from the table; account exposure strip
Data sources MetaTrader 5, Yahoo Finance, Binance, OANDA, Alpaca, CSV files and Rithmic futures data (optional async_rithmic package), switched while the terminal runs (details)
Paper trading on live prices Practise on any data source: a paper account fills your orders at its live bid/ask, and SL/TP trigger on wicks
Simulated market Bates-model prices (stochastic volatility + jumps) with an intraday rhythm, consistent across timeframes and reproducible from a seed; paper account with margin checks and SL/TP that trigger on wicks
Learning Center Bilingual (English / العربية) guide to the terminal, the quant models (GBM, MJD, Bates, GEX), trade setups and keyboard shortcuts
Desktop app Optional Electron shell that starts the backend and opens the terminal in its own window
Safety Trading switch off by default and enforced by the server, local-only server, protection against cross-site requests and DNS rebinding, API keys never sent back to the browser

Options levels — from public option chains

Panel What it shows
Options Quant Gamma exposure (GEX), the gamma flip, call/put walls, max pain, expected move, IV skew, dealer scenarios and the level ladders
Options Cockpit The options view in Stress Lab

The terminal computes these itself for US500, USTEC, XAUUSD, XAGUSD and XTIUSD from the SPY, QQQ, GLD, SLV and USO option chains (see How the Options Quant page is built). Three parts still need data the repository does not produce: the Heat Map surface, the Retail Trap card and the Vanna/Charm card.

Data sources

Prices can come from seven sources. Pick one in Settings ⚙ → DATA SOURCES while the terminal runs (SETTINGS, TEST, USE), or start with python launcher.py --provider <type>. MetaTrader 5 and OANDA also take your orders. The other sources only deliver prices, so they trade on the paper account, which fills orders at that source's live bid and ask.

Source --provider Markets Prices Needs Orders go to
MetaTrader 5 mt5 Everything your broker offers Real time Windows, the MT5 terminal, a broker account (demo is fine) MT5
OANDA oanda Forex; gold, silver, platinum, oil, natural gas, copper, corn, wheat, soybeans, sugar; index CFDs Real time A free practice account and an API token OANDA
Yahoo Finance yahoo Forex; metals, energy, grain, soft and livestock futures; indices, stocks, ETFs, crypto Delayed for some markets (futures ≈ 10 min) Nothing: free, no account Paper account
Binance binance Crypto Real time Nothing: free, no account Paper account
Alpaca alpaca US stocks and ETFs (including GLD, SLV, USO, UNG, CPER), crypto Real time (IEX) A free account and an API key Paper account
CSV files csv Anything you have bars for From the files Files in a folder Paper account
Simulated market sim The default tickers, with synthetic prices Generated Nothing, works offline The simulator's paper account

When you switch, the charts reload, and the cones and bands are rebuilt from the new source's daily bars within a minute or two. The header shows the active source: MT5 ● LINKED, FEED ● YAHOO, FEED ● BINANCE and so on.

Setting up each source

Yahoo Finance. There is nothing to set up. Quotes refresh every 30 seconds (Quote refresh; lower it at the risk of Yahoo's rate limit). A quote is the last trade price, so paper orders on Yahoo Finance fill without a spread. Metals, energy and US indices are quoted as front-month futures (gold is GC=F), so their levels differ slightly from spot or CFD prices. History goes back about 7 days for 1-minute bars, 60 days for 5- to 30-minute bars, 2 years for hourly bars and decades for daily bars.

Binance. There is nothing to set up. BTCUSD is read as BTCUSDT (change Quote asset for USDC or FDUSD). If api.binance.com is not reachable where you are, as in the US, pick https://api.binance.us or the market-data-only https://data-api.binance.vision as API address.

OANDA:

  1. Open a free practice account at oanda.com.
  2. In the OANDA hub, open My Account → Manage API Access and generate a token.
  3. In Settings ⚙ → DATA SOURCES → OANDA → SETTINGS, choose the practice environment, paste the token, and leave Account ID empty to use your first account. Then press SAVE, TEST and USE.

Orders then go to that OANDA account once LIVE TRADING is on. Daily bars start at the 17:00 New York roll, as on MetaTrader 5 forex brokers.

Alpaca:

  1. Create a free account at alpaca.markets. A paper-only account is enough.
  2. In the dashboard, generate an API key and copy the key ID and the secret. The secret is shown only once.
  3. In Settings ⚙ → DATA SOURCES → Alpaca → SETTINGS, paste both. Feed iex is free; sip covers every US exchange and needs a paid plan. Then press SAVE, TEST and USE.

CSV files. Put the files in the CSV folder, which by default is csv inside the AppData: folder. Name them XAUUSD_M15.csv, XAUUSD_D1.csv, or just XAUUSD.csv. The columns are found by name:

  • date and time, in one column or two;
  • open, high, low and close;
  • volume, which is optional.

Comma, semicolon and tab delimiters, most date formats and Unix time are recognized, so exports from MetaTrader 5, TradingView or NinjaTrader work as they are. A missing timeframe is built from a shorter one (H1 bars from M15, for example). For cones and bands, provide at least about a year of daily bars. The ticker must also be in your UNIVERSE list.

Keys without the Settings panel. The terminal also reads the environment variables OANDA_TOKEN, OANDA_ACCOUNT_ID, APCA_API_KEY_ID and APCA_API_SECRET_KEY. Keys stay on your computer, in user_config.json in the AppData: folder. They are sent only to that source's own API, and the Settings panel shows them as ••••••.

Ticker names

The terminal uses one name per instrument, and each source translates it:

Terminal Yahoo Finance OANDA Binance Alpaca
EURUSD EURUSD=X EUR_USD — —
XAUUSD GC=F XAU_USD — —
XTIUSD CL=F WTICO_USD — —
NATGAS NG=F NATGAS_USD — —
CORN ZC=F CORN_USD — —
US500 ES=F SPX500_USD — —
BTCUSD BTC-USD BTC_USD BTCUSDT BTC/USD

Add commodities under Settings ⚙ → UNIVERSE with the CMDTY class. Yahoo Finance knows XAGUSD, XPTUSD, XBRUSD, COPPER, WHEAT, SOYBEAN, COFFEE, SUGAR, COCOA and COTTON as they are, and OANDA all of them but coffee, cocoa and cotton. A ticker that stays empty is not known to that source under that name: set its broker symbol in UNIVERSE to the source's own name, for example ZO=F for oats on Yahoo Finance. For a mapping that applies to one source only, add it to providers.accounts.<id>.symbols in user_config.json, for example "symbols": {"US500": "SPX500_USD"}.

The paper account on live prices

  • It starts with 100,000 USD at 1:100 leverage. Both can be changed in the execution-account box of DATA SOURCES, and RESET starts the account over.
  • It supports market, limit and stop orders, stop loss and take profit, partial closes and margin checks, with P&L converted to USD. The account is saved in paper_account.json.
  • Stop loss and take profit trigger on wicks. The terminal checks every quote it saw and the source's 1-minute bars, but never a price from before the order existed.
  • As with every account, orders wait for Settings ⚙ → TRADING → LIVE TRADING.

Screenshots

Order ticket with stop loss, take profit and a risk/reward preview Learning Center showing the project guide in Arabic
Order ticket with a live risk / reward preview (paper account) The Learning Center in Arabic
Volatility cones with drift-based take-profit, entry and stop-loss levels Two charts, AUDUSD and EURUSD, with volatility cones and bands
Volatility cones with drift-based TP / entry / SL levels Cones and bands on two charts

The last two images were rendered from pre-computed analysis data, which is not part of this repository.

How the cones are computed

A few seconds after start-up, and again whenever a new daily bar appears, the terminal rebuilds every instrument's cones from the active feed's daily bars (backend/local_analysis/):

  • Anchors: each cone starts at the open of its period — the trading day, the week (Monday), the month or the quarter — and runs to the end of that period. The chart also shows the previous periods' cones, so you can see how past forecasts played out.
  • No look-ahead: a cone's parameters come only from the ~250 daily returns before its anchor.
  • Models: GBM (normal returns, closed form); MJD (GBM plus Poisson jumps, fitted from the days that stand out against the volatility of their time; exact mixture quantiles); Bates (Heston stochastic variance plus the same jumps, 25,000 Monte Carlo paths with a fixed seed, so results are reproducible).
  • Lines: the median and the ±1, ±2 and ±3 standard-deviation quantiles (p15.9 / p84.1, p2.3 / p97.7, p0.13 / p99.87), with a zero expected log return.
  • Trading time: time advances only on the weekdays the instrument actually trades, read from its own bars, so a cone does not widen over a closed weekend.

The results are written to the analysis cache as {TICKER}_cones.json and {TICKER}_quarterly_cones.json (eight quarters: the chart shows four, the REPLAY page all eight), and the past weekly and monthly cones the REPLAY page plays back as {TICKER}_historical_cones.json. A file another producer wrote is never overwritten. Turn the cones on from the ▾ next to CONES (and BATES) on a chart; MK_LOCAL_ANALYSIS=0 turns the computation off.

How the bands are computed

A cone is one forecast per period. The bands start again every trading day, so they follow the price:

  • Anchor: every trading day's open. The forecast uses only the returns before it.
  • Width: the volatility of the last 60 trading days, leaving out jump days. The tails keep the cones' long-run shape: MJD's jumps, and Bates' long-run variance, vol of vol and correlation, all from ~250 days. So the bands react to a change in volatility within weeks.
  • History: one point per day, half a trading day after the open, at that horizon's ±1/2/3-sigma quantiles. The chart joins and smooths these points into the envelope.
  • Forward fan: the newest day is projected 20 trading days ahead.
  • Centre line: the anchor price, shared by the three models, which all have a zero expected log return.

They are written as {TICKER}_bands.json (chart, Pulse Room), and every day's forward fan as {TICKER}_bands_replay.json for the REPLAY page. Turn them on with BANDS on a chart, or the B key, and pick GBM, MJD or BATES from its ▾.

How the regimes are computed

The REGIMES strip above the chart has two readings, both computed by the terminal (backend/local_analysis/regimes.py):

  • MICRO, the instrument's own regime. A hidden Markov model reads each day's return and the realized volatility of the last 5 days, from the active feed's daily bars.
    • It uses 2 to 4 states, chosen by BIC and ordered from calm to stressed by their volatility.
    • Walk-forward, no look-ahead: the model is refitted every month on up to four years of history, and each day is labelled with the filtered probabilities of a model fitted before it. The shading shows what the strip would have shown on each day.
    • The pill shows the regime expected for today, from the last completed day, and its probability.
  • ⚠ IMMINENT, the 15-minute volatility alarm.
    • The last 2,000 15-minute returns are divided by the usual size of moves at that time of day, so the New York open is not mistaken for a regime.
    • A two-state model (normal / burst) then forecasts the volatility of the next two hours.
    • IMMINENT appears when that forecast is at least 1.5 times the usual for those hours. Otherwise the chip shows QUIET and the minutes since the last alarm.
  • MACRO, the economy's quadrant.
    • Growth is the 3-month change of the S&P 500.
    • Inflation is the 3-month change of the 10-year breakeven rate and of WTI crude.
    • Each is measured against its typical size over the last three years.
    • A four-state HMM pinned to the corners turns them into GOLDILOCKS (growth up, inflation down), REFLATION (both up), STAGFLATION (growth down, inflation up) or DEFLATION (both down).
    • The series come from FRED, free and without a key, so MACRO needs internet access to fred.stlouisfed.org.

The results are written as {TICKER}_regime.json and GLOBAL_macro_regime.json. The daily part is rebuilt once per trading day and the alarm with every new 15-minute bar. Turn the shading on with MACRO or MICRO above the chart; the ▾ next to each sets the opacity, hides states or paints the candles.

How the Fundamental page is built

The FUNDAMENTAL page is built by the terminal from free public data (backend/local_analysis/fundamental.py, cot.py):

Card Data Notes
Macro regime The MACRO quadrant (S&P 500, 10-year breakeven, WTI from FRED) Probability, the label 30 days ago, and a 12-month trail on the growth/inflation map
Net liquidity Fed balance sheet (WALCL) − Treasury General Account (WTREGEN) − reverse repo (RRPONTSYD) Weekly, with the Fed's rate changes (DFEDTARU) marked
Yield curve, 2s10s 2/5/10/30-year yields (DGS2…DGS30), now and 3 months ago Also the 10-year real yield (DFII10) and fed funds (DFF)
VIX VIXCLS Above 25 is fear, below 18 is calm
COT positioning CFTC Commitments of Traders, speculators' net position A z-score against the last three years; ±1.5 is crowded
Asset scorecard Your data source's daily bars −100…+100: the regime's rule of thumb for the asset class (40%), the price vs its 200-day average (30%), the 3-month change (30%)

It needs internet access to fred.stlouisfed.org and publicreporting.cftc.gov. Downloads are cached (FRED for 6 hours, COT for 12), so the page keeps the latest data while offline, and a card whose data was never fetched is left out. The asset scores are refreshed every hour.

How the Money Flow page is built

The MONEY FLOW page is rebuilt every hour from daily ETF prices and volumes on Yahoo Finance (backend/local_analysis/money_flow.py):

  • CAPITAL FLOWS: seven asset classes, each read through representative ETFs:

    • equities: SPY, QQQ, IWM, EFA, EEM;
    • bonds: TLT, IEF, AGG, LQD, HYG;
    • commodities: GLD, SLV, USO, DBC, DBA, CPER;
    • crypto: IBIT, FBTC, ETHA;
    • T-bills: SHV, BIL, SGOV;
    • the dollar: UUP;
    • volatility hedges: VXX, UVXY.

    A class's net flow is a proxy, not reported fund flows: each day's return times the dollars traded that day, summed over 4 weeks (MACRO) or 1 week (MICRO). The arcs run from the classes under selling pressure to the ones being bought.

  • SECTORS: the 11 SPDR sector ETFs, their 1- and 4-week change.

  • GLOBAL: country ETFs on the map, with arcs from the three weakest markets to the three strongest.

A ticker that can't be downloaded is left out rather than shown as flat. The page needs internet access to query1.finance.yahoo.com, and the world map's shapes come from cdn.jsdelivr.net.

How the Options Quant page is built

Every 15 minutes while New York trades (every 2 hours otherwise) the terminal downloads the option chains of SPY, QQQ, GLD, SLV and USO from Yahoo Finance (yfinance, no key; Yahoo's option quotes are about 15 minutes delayed) and writes {T}_options.json for US500, USTEC, XAUUSD, XAGUSD and XTIUSD, when they are in your universe and your data source has a price for them.

  • Expiries: today's (0-DTE, until 16:00 New York), the next eight after today (the swing view) and the next two monthly expiries.
  • Implied vols: Yahoo's, cleaned of out-of-range values and of its failed solves, then one smile per expiry from the out-of-the-money side.
  • GEX: dealers are assumed long the calls and short the puts customers trade. GEX is in dollars per 1% move. The flip is the price where total swing GEX changes sign. The call wall is the strike above spot with the most call gamma, the put wall the strike below spot with the most put gamma, and HVL the strike with the most gamma overall. Max pain and the expected move (the ATM straddle) use the first expiry after today.
  • Dealer delta and scenarios are in shares of the ETF: how much the dealers' hedge would change if price moved 0.5% or 1% either way. The asymmetry is their selling over their buying across ±1%.
  • Structural Integrity and Volatility Gravity use the UI's own formulas (frontend/src/lib/options_0dte_derivations.js), ported to Python.
  • Price space: every level is multiplied by your feed's price over the ETF's price, both taken at the time of the ETF quote, so the lines sit on your chart (for example SPY 680 → US500 6,800).

It needs internet access to query1.finance.yahoo.com and query2.finance.yahoo.com. When a download fails the last file stays. These levels describe positioning; they are not trade signals on their own.

How it works

flowchart LR
    subgraph feeds["Market data"]
        MT5["MetaTrader 5"]
        NET["Yahoo Finance · Binance<br/>OANDA · Alpaca"]
        CSV["CSV files"]
        RITH["Rithmic"]
        SIM["Simulated market<br/>(Bates model)"]
    end
    CACHE[("Analysis cache<br/>JSON files")]
    feeds --> API["Python backend<br/>FastAPI · 127.0.0.1:8502"]
    API -- "daily bars" --> ENGINE["Local analysis<br/>cones · bands · regimes"]
    FRED["FRED · CFTC · ETF prices<br/>(macro data)"] --> ENGINE
    ENGINE --> CACHE
    CACHE --> API
    API -- "REST + WebSocket" --> UI["React UI<br/>frontend/"]
    UI --> BROWSER["Your browser"]
    UI --> ELECTRON["Electron window"]
Loading

The backend talks to the active data source through a common provider interface, sends orders to the execution account (MetaTrader 5, OANDA or a paper account), streams prices to the UI over WebSockets, computes the cones, bands and regimes in the background and serves the analysis cache over REST. The UI is a React 18 app; its production build is committed in frontend_build/, so running the terminal never requires Node.js.

QuantumTerminal/
├── backend/                 Python backend (FastAPI)
│   ├── launcher.py          entry point: config, UI serving, command-line options
│   ├── data_server.py       REST API + WebSocket price streams
│   ├── providers/           data sources (MT5, Yahoo, Binance, OANDA, Alpaca, CSV, Rithmic), simulator, paper account
│   ├── local_analysis/      cones, bands, regimes, Replay files, the Fundamental and Money Flow pages
│   ├── security.py          local-only API guard (cross-site / DNS-rebinding protection)
│   └── tests/               pytest suite — runs without a broker
├── frontend/                React source of the UI (Create React App)
├── frontend_build/          production build served by the backend
├── electron_shell/          optional desktop app
└── screenshots/

Configuration

Option Environment variable What it does
--sim MK_PROVIDER=sim Simulated market data + paper trading account
--provider yahoo MK_PROVIDER=yahoo Data source for this run: mt5, sim, yahoo, binance, oanda, alpaca or csv
OANDA_TOKEN, OANDA_ACCOUNT_ID OANDA API token and account, instead of entering them in Settings
APCA_API_KEY_ID, APCA_API_SECRET_KEY Alpaca API key pair, instead of entering it in Settings
--port 8600 MK_PORT HTTP port (default 8502)
--no-browser MK_NO_BROWSER=1 Don't open the system browser
MK_UPDATE_CHECK=0 Don't check GitHub for a newer release (the answer is cached for 10 minutes)
MK_LOCAL_ANALYSIS=0 Don't compute the cones, bands and regimes locally
MK_FRONTEND_DIR Serve a different UI build
MK_ALLOWED_HOSTS Extra host names allowed to reach the API, e.g. a LAN address (* disables the check)
MK_CORS_ORIGINS Extra browser origins allowed to call the API (default: http://localhost:3000)

Your settings are saved in user_config.json in the folder printed as AppData: when the terminal starts. Each source's settings are under providers.accounts.<id> (yahoo_default, binance_default, oanda_default, alpaca_default, csv_default); the simulator's seed, starting balance and leverage are under sim_default, and the paper account's under paper_default.

Development

  • Backend — cd backend && pip install -r requirements-dev.txt && python -m pytest. The suite runs the real API end-to-end on the simulated market, so it needs no broker.
  • UI — the React source lives in frontend/: npm ci && npm start gives hot reload against a running backend. Run npm run build and commit frontend_build/ together with your source change.
  • CI runs the tests on Linux and Windows with Python 3.10 and 3.12, rebuilds the UI to check that frontend_build/ matches the source, and checks the Electron shell.

See CONTRIBUTING.md for the workflow, and CHANGELOG.md for what changed in each release.

Security

The backend listens on 127.0.0.1 only. It rejects requests from other websites (Origin and Sec-Fetch-Site checks, WebSocket handshakes included) and DNS-rebinding attempts (Host header allow-list), so a web page you visit cannot place orders or read data through your running terminal. Orders are blocked until you switch trading on, and the server refuses them too. API tokens and secrets are sent only to their data source and never back to the browser. Calls into MetaTrader 5 are serialized.

Found a vulnerability? Please report it privately; see SECURITY.md.

Roadmap

  • Local quant engine, next steps — the cones, bands, regimes, the Fundamental and Money Flow pages are computed locally; options levels are next, so every panel works without a data server.
  • More brokers — Interactive Brokers, and orders on Alpaca, through the provider interface. A data source is about 100 lines: see providers/binance_provider.py.
  • Translations — the Learning Center already supports English and Arabic; the rest of the UI is English for now.
  • Packaged releases — signed desktop installers.

Contributions are welcome — see CONTRIBUTING.md.

Disclaimer

Quantum Terminal is for education and analysis. Trading financial markets carries a high risk of loss and is not suitable for everyone. The models are probabilistic, not predictive. Prices from the simulated market (--sim) are synthetic and must never be used as market data. Prices from Yahoo Finance may be delayed.

الشرح بالعربية

ما هي Quantum Terminal؟

منصة تداول مجانية ومفتوحة المصدر تعمل على جهازك أنت. معظم منصات الأفراد مبنية حول مؤشرات متأخرة مثل RSI وMACD، أما Quantum Terminal فترسم فوق الشموع الحية المدى الذي يُرجَّح إحصائياً أن يتحرك فيه السعر: نطاقات توقّع (volatility cones) من نماذج GBM وMJD وBates، تحسبها المنصة بنفسها من بيانات السوق. ومعها شارت حي، وإدخال أوامر مع وقف الخسارة وجني الربح، وإدارة للصفقات، وسوق محاكى كامل بحساب تداول ورقي لتجرّب كل شيء دون أي وسيط ودون أي مخاطرة. لا تسجيل حساب، ولا مفتاح ترخيص، ولا اشتراك مدفوع.

أبرز المزايا

  • الـ volatility cones: يومية وأسبوعية وشهرية وفصلية بثلاثة نماذج (GBM وMJD وBates)، تُحسب من الشموع اليومية لمصدر البيانات نفسه، دون النظر إلى المستقبل، وتتحدّث تلقائياً مع كل شمعة يومية جديدة.
  • الـ bands الاحتمالية: تبدأ من جديد عند افتتاح كل يوم تداول ويُحسب اتساعها من آخر 60 يوم تداول، فتتحرك مع السعر، مع توقّع لـ 20 يوم تداول قادمة. وصفحة REPLAY تعيد عرض الـ cones والـ bands السابقة أمام حركة السعر شمعة بشمعة.
  • الـ Regimes: حالة التقلب لكل أداة من الهدوء (calm) إلى الضغط (stressed)، وتنبيه IMMINENT عند توقّع انفجار في التقلب خلال ساعتين، وربع الاقتصاد الكلي (GOLDILOCKS وREFLATION وSTAGFLATION وDEFLATION) من بيانات FRED، وكلها دون النظر إلى المستقبل.
  • صفحة Fundamental: حالة الاقتصاد الكلي، وصافي سيولة الفيدرالي، ومنحنى العوائد، ومؤشر VIX، ومراكز المضاربين في العقود الآجلة (CFTC COT)، ودرجة لكل أداة، من بيانات عامة مجانية.
  • صفحة Money Flow: دوران الأموال بين فئات الأصول والقطاعات الأمريكية والدول، من أسعار صناديق ETF وأحجام تداولها.
  • شارت حي: من M1 إلى D1، عدة شارتات في تبويبات، أدوات رسم، قوالب، وساعة عالمية بجلسات التداول.
  • الأوامر والصفقات: أوامر market وlimit وstop مع SL/TP، ومعاينة فورية للمخاطرة والعائد، وسحب خطوط SL/TP على الشارت.
  • السوق المحاكى: أسعار من نموذج Bates متسقة بين كل الأطر الزمنية، وحساب ورقي برصيد 100,000 دولار.
  • مصادر البيانات: MetaTrader 5 وYahoo Finance وBinance وOANDA وAlpaca وملفات CSV، مع التبديل بينها أثناء عمل المنصة، وحساب ورقي يتداول على الأسعار الحية لأي منها.
  • السلع: الذهب والفضة والبلاتين والنفط والغاز الطبيعي والنحاس، والحبوب والسكر والقهوة والكاكاو والقطن، عبر MetaTrader 5 أو OANDA أو عقود futures من Yahoo Finance.
  • مركز التعلم: دليل كامل بالعربية والإنجليزية داخل المنصة.
  • الأمان: السيرفر محلي فقط، ولا تستطيع أي صفحة ويب تزورها إرسال أوامر عبر منصتك، ومفاتيح الـ API لا تغادر جهازك إلا إلى مصدرها.

تحسب المنصة مستويات Options Quant بنفسها من سلاسل خيارات SPY وQQQ وGLD وSLV وUSO العامة، ما عدا الخريطة الحرارية وبطاقتي Retail Trap وVanna/Charm. التفاصيل في النسخة العربية الكاملة.

المتطلبات

المطلوب ملاحظات
نظام التشغيل Windows 10 أو 11، أو macOS، أو Linux بنظام 64 بت MetaTrader 5 يحتاج Windows، أما السوق المحاكى فيعمل على الأنظمة الثلاثة
Python الإصدار 3.10 أو أحدث، بنسخة 64 بت جُرّبت المنصة على 3.10 و3.12 و3.13. على Windows فعّل خيار Add python.exe to PATH أثناء التثبيت
الذاكرة ذاكرة بسعة 4 غيغابايت الـ backend نفسه يستهلك نحو 150 ميغابايت
مساحة القرص نحو 1 غيغابايت فارغة حزم Python تأخذ نحو 500 ميغابايت
المتصفح نسخة حديثة من Chrome أو Edge أو Firefox أو Safari أو استخدم تطبيق سطح المكتب بدلاً منه
الإنترنت للتثبيت، ثم لمصادر البيانات الحية وبيانات الاقتصاد الكلي (FRED وCFTC وYahoo Finance) والتحقق من التحديثات السوق المحاكى وملفات CSV تعمل دون إنترنت
MetaTrader 5 ‏(اختياري) منصة MT5 من وسيطك، مع تسجيل الدخول إلى حساب (الحساب التجريبي يكفي) على Windows فقط، وهي إحدى طرق الحصول على أسعار حقيقية وإرسال أوامر حية، والطرق الأخرى في مصادر البيانات
Node.js ‏(اختياري) الإصدار 20 أو أحدث فقط لتطبيق سطح المكتب أو لإعادة بناء الواجهة
Git ‏(اختياري) أي إصدار أو حمّل الكود كملف ZIP

التثبيت والتشغيل خطوة بخطوة

1. ثبّت Python

  • على Windows: حمّل مثبّت نسخة 64 بت من python.org، وفعّل خيار Add python.exe to PATH في الشاشة الأولى. ثم افتح نافذة Command Prompt جديدة وتأكد أن الأمر python --version يُظهر 3.10 أو أحدث.
  • على macOS: استخدم المثبّت من python.org أو الأمر brew install python@3.12. وعلى macOS وLinux اكتب python3 أينما ذُكر python في هذا الدليل.
  • على Linux: ثبّت حزم توزيعتك، فعلى Debian وUbuntu مثلاً نفّذ الأمر sudo apt install python3 python3-venv python3-pip في نافذة الأوامر.

2. احصل على الكود

git clone https://github.com/wJubran/QuantumTerminal.git
cd QuantumTerminal

بدون Git: من صفحة المستودع اضغط Code ← Download ZIP، وفُكّ الضغط، ثم افتح نافذة الأوامر داخل المجلد الناتج، وقد يكون اسمه QuantumTerminal-main بعد فك الضغط.

3. أنشئ بيئة افتراضية (مُستحسن)

البيئة الافتراضية (virtual environment) تعزل حزم المنصة عن بقية النظام. أنشئها داخل مجلد المشروع، فهناك يبحث عنها تطبيق سطح المكتب أيضاً:

:: Windows (Command Prompt)
python -m venv .venv
.venv\Scripts\activate
# macOS / Linux
python3 -m venv .venv
source .venv/bin/activate

سيبدأ سطر الأوامر الآن بـ (.venv). فعّل البيئة من جديد في كل نافذة أوامر جديدة قبل تشغيل المنصة. وإذا رفض PowerShell تشغيل سكربت التفعيل، فنفّذ مرة واحدة الأمر Set-ExecutionPolicy -Scope CurrentUser RemoteSigned، أو استخدم نافذة Command Prompt بدلاً منه.

4. ثبّت المكتبات المطلوبة

cd backend
pip install -r requirements.txt

يستغرق التثبيت الأول بضع دقائق. حزمة MetaTrader5 تُثبَّت على Windows فقط، لأن شركة MetaQuotes لا تنشرها إلا لنظام Windows وحده.

5. شغّل المنصة

على السوق المحاكى (أي نظام، دون حساب لدى وسيط):

python launcher.py --sim

على أسعار حية، بدون وسيط (أي نظام، ويكفي الاتصال بالإنترنت):

python launcher.py --provider yahoo      # forex, metals, energy, grains, indices, crypto
python launcher.py --provider binance    # crypto, in real time

بعدها تذهب الأوامر إلى حساب ورقي ينفّذها بتلك الأسعار الحية. أما OANDA وAlpaca وملفات CSV فتُعدّ من الإعدادات ⚙ ← DATA SOURCES، والتفاصيل في مصادر البيانات.

مع MetaTrader 5 (على Windows):

  1. ثبّت MetaTrader 5 من وسيطك وسجّل الدخول إلى حسابك (الحساب التجريبي يكفي).
  2. لتتمكن من إرسال الأوامر من Quantum Terminal، فعّل Algo Trading في MetaTrader 5 من زره في شريط الأدوات، أو من القائمة Tools → Options → Expert Advisors بتفعيل الخيار Allow algorithmic trading فيها.
  3. اترك MetaTrader 5 مفتوحاً وشغّل الأمر: python launcher.py
  4. إذا لم تعثر المنصة على MetaTrader 5 تلقائياً، فاضبطه من الإعدادات ⚙ ← PROVIDERS: إما TERMINAL PATH (مسار الملف terminal64.exe) أو LOGIN CREDENTIALS (السيرفر ورقم الحساب وكلمة المرور).

يطبع المشغّل عنوان السيرفر وأماكن ملف الإعدادات والبيانات (AppData:) وملف السجل، ثم يفتح العنوان http://127.0.0.1:8502 في متصفحك. إذا لم يُفتح المتصفح فافتح العنوان بنفسك. واترك نافذة المشغّل مفتوحة طوال استخدامك للمنصة.

6. أول خطوات داخل المنصة

  • الـ cones: على الشارت افتح السهم ▾ بجانب CONES واختر DAILY أو WEEKLY أو MONTHLY أو QUARTERLY، ومنه أيضاً BATES لعرض نموذج Bates. في المرة الأولى تكون جاهزة بعد التشغيل بنحو 30 ثانية.
  • التداول: الأوامر محظورة إلى أن تفعّل الإعدادات ⚙ ← TRADING ← LIVE TRADING. وفي المحاكي، وفي المصادر التي تقدّم الأسعار فقط، يفتح هذا المفتاح الحساب الورقي فقط (100,000 دولار برافعة 1:100).
  • مصدر البيانات: من الإعدادات ⚙ ← DATA SOURCES تنتقل بين MetaTrader 5 وYahoo Finance وBinance وOANDA وAlpaca وملفات CSV والمحاكي أثناء عمل المنصة، وتتذكر المنصة اختيارك.
  • مركز التعلم: تبويب LEARN يشرح كل صفحة وكل نموذج، بالعربية والإنجليزية.

7. الإيقاف وإعادة التشغيل والتحديث

  • الإيقاف: اضغط Ctrl+C في نافذة المشغّل، أو أغلقها.
  • إعادة التشغيل: افتح نافذة أوامر داخل مجلد المشروع، وفعّل البيئة الافتراضية (الخطوة 3)، ثم نفّذ cd backend وpython launcher.py (وأضف --sim للسوق المحاكى).
  • التحديث: داخل مجلد المشروع نفّذ git pull، ثم pip install -r backend/requirements.txt والبيئة الافتراضية مفعّلة. تبقى إعداداتك والحساب الورقي كما هي، لأنها محفوظة في مجلد AppData: لا في مجلد المشروع.
  • إزالة المنصة: احذف مجلد المشروع ومجلد AppData: الذي يظهر عند التشغيل.

تطبيق سطح المكتب (اختياري)

بعد تثبيت Node.js بالإصدار 20 أو أحدث:

cd electron_shell
npm install
npm start              # or: npm run start:sim

يشغّل التطبيق الـ backend تلقائياً، مستخدماً البيئة .venv أو venv في مجلد المشروع إن وُجدت، وإلا فـ Python المثبّت على الجهاز، ويعرض المنصة في نافذة مستقلة. ويحفظ إعداداته في مجلد خاص به اسمه QuantumTerminal-v2، منفصل عن إعدادات نسخة المتصفح.

حل المشكلات الشائعة

المشكلة الحل
رسالة 'python' is not recognized على Windows شغّل مثبّت Python مجدداً وفعّل Add python.exe to PATH، أو اكتب py بدلاً من python
رسالة error while attempting to bind on address المنفذ 8502 مستخدم، غالباً بنسخة أخرى من المنصة. أغلقها، أو شغّل المنصة على منفذ آخر بإضافة --port 8600 إلى الأمر
لا توجد حزمة MetaTrader5 على macOS أو Linux هذا متوقع، فهي متوفرة لنظام Windows فقط. استخدم الخيار --sim بدلاً منها
رسالة MT5 initialize() failed في السجل تأكد أن MetaTrader 5 مفتوح ومسجَّل الدخول، واضبط TERMINAL PATH من الإعدادات ⚙ ← PROVIDERS
MetaTrader 5 يرفض الأوامر لأن التداول الآلي معطّل فعّل Algo Trading في MetaTrader 5
لم يُفتح المتصفح، أو بقيت الصفحة فارغة افتح http://127.0.0.1:8502 بنفسك، وراجع ملف السجل الذي طبعه المشغّل
لا تظهر الـ cones على الشارت انتظر نحو 30 ثانية بعد التشغيل الأول، ثم فعّلها من السهم ▾ بجانب CONES
رمز يبقى فارغاً على Yahoo Finance أو OANDA أو Binance أو Alpaca المصدر يعرفه باسم آخر: اضبط له broker symbol من الإعدادات ⚙ ← UNIVERSE (راجع أسماء الرموز)
زر TEST في DATA SOURCES يُظهر الخطأ 401 أو 403 الـ token أو المفتاح خاطئ أو منتهي الصلاحية، أو أنك تستخدم token من practice على live (أو العكس)
زر TEST لا يصل إلى المصدر اسمح بعنوانه في جدار الحماية أو الـ proxy: query1.finance.yahoo.com أو api.binance.com أو api-fxpractice.oanda.com (وapi-fxtrade.oanda.com للحساب الحقيقي) أو data.alpaca.markets

مصادر البيانات

اختر مصدر الأسعار من الإعدادات ⚙ ← DATA SOURCES أثناء عمل المنصة (بالأزرار SETTINGS وTEST وUSE)، أو بالخيار --provider عند التشغيل. يستقبل MetaTrader 5 وOANDA أوامرك أيضاً، أما بقية المصادر فتُنفَّذ أوامرها على حساب ورقي (paper account) بأسعارها الحية.

المصدر --provider الأسواق الأسعار المطلوب تذهب الأوامر إلى
MetaTrader 5 mt5 كل ما يقدّمه وسيطك لحظية Windows ومنصة MT5 وحساب لدى وسيط (الحساب التجريبي يكفي) MT5
OANDA oanda العملات؛ والذهب والفضة والبلاتين والنفط والغاز الطبيعي والنحاس والذرة والقمح وفول الصويا والسكر؛ ومؤشرات CFD لحظية حساب practice مجاني و API token OANDA
Yahoo Finance yahoo العملات؛ وعقود futures للمعادن والطاقة والحبوب والسلع الطرية والمواشي؛ والمؤشرات والأسهم وصناديق ETF والعملات الرقمية متأخرة في بعض الأسواق (نحو 10 دقائق للـ futures) لا شيء: مجاني وبدون حساب الحساب الورقي
Binance binance العملات الرقمية لحظية لا شيء: مجاني وبدون حساب الحساب الورقي
Alpaca alpaca الأسهم وصناديق ETF الأمريكية (ومنها GLD وSLV وUSO وUNG وCPER)، والعملات الرقمية لحظية (IEX) حساب مجاني و API key الحساب الورقي
ملفات CSV csv أي أداة تملك شموعها من الملفات ملفات في مجلد الحساب الورقي
السوق المحاكى sim الرموز الافتراضية بأسعار اصطناعية مولَّدة لا شيء، ويعمل دون إنترنت الحساب الورقي للمحاكي
  • Yahoo Finance وBinance: لا يحتاجان أي إعداد ولا أي حساب.
  • OANDA: افتح حساب practice مجانياً في oanda.com، وأنشئ token من My Account ← Manage API Access، ثم الصقه في DATA SOURCES ← OANDA ← SETTINGS واضغط SAVE ثم TEST ثم USE.
  • Alpaca: أنشئ حساباً مجانياً في alpaca.markets، وأنشئ API key من لوحة التحكم، ثم الصق الـ key ID والـ secret في DATA SOURCES ← Alpaca ← SETTINGS واضغط SAVE ثم TEST ثم USE.
  • ملفات CSV: ضع ملفات مثل XAUUSD_D1.csv في المجلد csv داخل مجلد AppData:، بأعمدة التاريخ والوقت وopen وhigh وlow وclose.
  • أسماء الرموز: كل مصدر يترجم أسماء المنصة إلى أسمائه، فـ XAUUSD هو GC=F في Yahoo Finance وXAU_USD في OANDA. والرمز الذي يبقى فارغاً اضبط له broker symbol من الإعدادات ⚙ ← UNIVERSE.
  • المفاتيح: تبقى على جهازك ولا تُرسَل إلا إلى المصدر نفسه، ولوحة الإعدادات تعرضها بالشكل ••••••.

الشرح الكامل لكل مصدر، مع جدول أسماء الرموز والحساب الورقي، في النسخة العربية الكاملة، وفي مقال مصادر البيانات: الربط والإعداد داخل مركز التعلم.

لشرح طريقة حساب الـ cones، والإعدادات، والتطوير، والأمان، راجع النسخة العربية الكاملة.

License

MIT © Quantum Terminal contributors

About

Free & open-source trading terminal: volatility cones, bands & options levels on live charts — MetaTrader 5 & Rithmic feeds, built-in simulated market, English & Arabic guide.

Topics

algorithmic-tradingday-tradingelectronfastapimetatrader5pythonrithmictradingtradovatevolatility

Resources

Readme
MIT license

Contributing

Contributing

Security policy

Security policy
Activity

Stars

31 stars

Watchers

3 watching

Forks

15 forks
Report repository

Releases

Packages

Contributors

Languages