Tasilab — Dashboard

Welcome to Tasilab. Let's run something.

Sign in to start logging experiments and paper trading on the Saudi Exchange.

Total experiments

—

across all types

Running

—

not yet sealed

Simulated trades

—

closed round-trips

Overall win rate

—

weighted across experiments

Currently running

No runs in flight

Start an experiment from your SDK and it will show up here until you call .end().

Highlights — top by each metric

No completed experiments yet

Highlights appear here once you've finished at least one run.

Recent experiments

View all →

Name	Type	Symbol	Status	Return	Created
Sign in to see your dashboard.

Every night you've answered today's question, with history's verdict once it seals.

Data delayed 15 minutes

Place an order

Symbol

Side

Buy

Sell

Order type

Market

Limit

Quantity

—

Trade value

—

Commission (0.155%)

—

VAT (15% of commission)

—

Total required

—

Paper wallet

—

SAR · starting balance 100,000

In positions

—

SAR deployed

Total PnL

—

SAR · 0.00%

Win rate

—

0W · 0L

Top symbols by P&L

Computed from your trade history

No trades yet

Once you place orders via the API, your top symbols will appear here grouped by realized P&L.

Trade history

Date	Symbol	Side	Quantity	Price	Value (SAR)	Commission	Status

Telegram notifications

Loading…

Edge Explorer

Ask one question at a time: pick a condition, pick a forward horizon, see how that slice of history compares to the unconditional baseline. No strategy, no fees, no P&L — just whether the edge exists before you commit to building a rule around it.

Symbol

History window

Condition type

Forward horizon

Day

Result

Finding

A historical pattern on one stock. Not a forecast, and not advice.

What was tested—

Times it happened—

Window searched—

Measured over—

For the technically minded

How the words above map to the statistics below.

Matched bars—

Total bars—

Match rate—

After it happened

Average move—

Median—

Ended higher—

Std dev—

Normally

Average move—

Median—

Ended higher—

Std dev—

Mean-return edge—

Win-rate edge—

p-value vs. baseline—

How big the moves were: after it happened vs all days

After it happened Normally (all days)

What this shows

How this stock behaved after the times this condition occurred, against how it behaved on every other day in the same window.

What this does not show

Cause, or what happens next. One search is one search: try enough conditions and some will look unusual by luck alone.

Run details & reproduction

Run ID—

Method version—

Data fingerprint—

—

Test a strategy on real TASI history

We start you with SAR 100,000, run your rule against real historical prices from the period you pick, apply TASI commission & VAT, and show what the account would be worth at the end.

Symbol

Period

All available

1 Year

3 Years

5 Years

10 Years

Strategy type

Indicator

Period

BUY when RSI is

SELL when RSI is

Risk management (optional)

Stop-loss % (exit if price drops this much below entry)

Take-profit % (exit if price rises this much above entry)

Result

—

One stock, one period. This is what happened, not what will happen.

Long only One stock All available cash per entry Commission, VAT and slippage applied No minimum commission

Starting capital

SAR 100,000

—

Final value

—

Return

—

Trades

—

Max drawdown

—

Worst fall from a peak

Equity curve

Your strategy

Buy & hold same stock

Trades

Date	Side	Price	Quantity	Value (SAR)	Reason

What this shows

What this exact rule would have done on this one stock over this one period, paying real TASI commission, VAT and slippage.

What this does not show

That the rule will work next year, on another stock, or after a minimum commission. One period is one sample, and a rule tuned until it looks good on that sample usually stops working on the next one.

Run details & reproduction

Run ID—

Method version—

Data fingerprint—

—

Your API key

No API key set. Sign up to get one.

Pass this key in the X-API-Key header with every request.

Quick start (Python SDK)

pip install tasilab

from tasilab import Tasilab

tasi = Tasilab(api_key="YOUR_API_KEY")

# 1. Pull historical bars (Tasilab caches them for you)

hist = tasi.get_historical("1120", "2024-01-01", "2025-12-31")

# 2. Run your strategy locally → list of trade dicts

trades = my_strategy(hist["bars"])

# 3. Log the run as a Tasilab experiment

with tasi.create_experiment(

name="MACD 12/26/9 on Al Rajhi",

symbol="1120",

start_date="2024-01-01", end_date="2025-12-31",

parameters={"fast": 12, "slow": 26, "signal": 9},

) as exp: # auto-seals on exit

exp.log_trades(trades)

exp.log_metrics({"total_return_sar": 439.64, "win_rate": 0.412})

Endpoints

GET /v1/auth/meCurrent user profile (id, email, api_key)

GET /v1/market/quote/{symbol}TASI quote (15-min delayed)

GET /v1/market/statusIndex value, open/closed

POST /v1/ordersPlace a market or limit order

GET /v1/ordersList orders

GET /v1/portfolioCash, positions, total return

GET /v1/portfolio/tradesTrade history

GET /v1/analytics/performanceWin rate, P&L, trade stats

Every plan starts you with SAR 100,000 of paper capital and real TASI market data. Pick the one that matches how much you want to explore.

Paid plans are launching soon — your Free plan access isn't changing.

Free

0 SAR/mo

.

Try the product, no card required

✓SAR 100,000 paper wallet (1 wallet)

✓Delayed TASI quotes (15-min lag)

✓Unlimited history (daily bars)

✓Market orders · max 5 open positions

✓3 strategy runs (lifetime)

✓5 backtests / month · 5 AI calls

✓60 API requests / minute

Most popular

Pro

129 SAR/mo

or 1,290 SAR/yr · 2 months free

Multi-wallet + AI

Everything in Free, plus:

✓TASI quotes, delayed 15 minutes

✓Unlimited history (daily + weekly)

✓Market + limit orders · stop-loss / TP

✓Unlimited wallets & positions

✓Unlimited strategy runs · unlimited backtests / mo

✓2 concurrent backtests · slippage simulation

✓50 AI analytics / month

✓Compare up to 3 strategies side-by-side

✓500 API / minute · 3 webhooks · 2 keys

✓Priority email (48h SLA)

Quant

399 SAR/mo

or 3,990 SAR/yr · 2 months free

The full parallel lab

Everything in Pro, plus:

✓5 concurrent backtests

✓Intraday bars (1m/5m)

✓WebSocket streaming · order book snapshot

✓STOP + OCO orders · configurable slippage

✓Unlimited side-by-side comparison · grid sweep

✓250 AI analytics / month (advanced model)

✓2,000 API / minute · 10 webhooks · 5 keys

✓Dedicated support (24h SLA · Telegram)

Team

999 SAR/mo

or 9,990 SAR/yr · 2 months free

Multi-seat workspace · shared library

Everything in Quant, per seat:

✓5 seats included · add-on seats available

✓Shared strategy library

✓10 concurrent backtests per org

✓250 AI / member / month (1,250 for 5 seats)

✓5,000 API / minute · 20 webhooks · 10 keys

✓Dedicated support · 99.5% SLA

✓For prop desks, university programs, small quant teams

Full feature comparison

Every limit, every gate, every feature flag — reads from the same source the matching engine and rate limiter use.

Loading…

Get in touch

📝 Contact form

📖 API docs

System status

API checking…

Database checking…

Market data checking…

Send us a message

Bug reports, feature requests, or anything else. We typically reply within a couple of days.

Your name (optional)

Email

Subject (optional)

Message

Regulation & advice

Is Tasilab a financial advisor?

No. Tasilab is a research, analytics, and paper-trading platform — a developer tool. It does not provide investment advice, recommendations, portfolio management, or any regulated financial service. Every analysis, backtest result, AI explanation, and exported report is educational content on historical data. Do not treat any output as a basis for real-money trading decisions. For investment advice, consult an advisor licensed by the Saudi Capital Market Authority.

Are you regulated by the Saudi Capital Market Authority (CMA)?

No, and we don't need to be. The Capital Market Law regulates licensed activities — placing real orders, custodying client money, providing personalised investment advice, managing portfolios. Tasilab does none of those. We simulate trades against real market data for research and education only. No real orders are routed; no client money is held; no recommendations are issued. See our disclaimer for the full position.

Can I use the AI explanations as a basis for real trades?

No. AI explanations describe historical statistical results in plain language. They are explicitly prohibited from recommending specific trades, suggesting buy/sell actions, or predicting future prices. The model is instructed to refuse personalised-advice questions and to direct you to a CMA-licensed advisor. If an AI response ever drifts into advisory tone, treat it as a system error and please report it.

What if I share a backtest publicly — could that be considered providing advice?

Sharing a backtest run via public link or social media is sharing your own research output, similar to publishing a blog post about a strategy idea. It's not regulated investment advice as long as you're not soliciting personalised investment relationships or charging for it. The export footer ("Educational simulation. Not investment advice.") makes the framing explicit. If you do plan to charge for advice or solicit clients, get a CMA license first.

Frequently asked questions

How do I place a paper trade?

Use the POST /v1/orders endpoint with your API key. Specify the symbol (e.g. 2222 for Aramco), side (BUY/SELL), order_type (MARKET/LIMIT), and quantity.

What are Saudi Exchange trading hours?

The Saudi Exchange (Tadawul) trades Sunday–Thursday, 10:00–15:00 AST. Orders submitted outside these hours are queued as PENDING and fill at market open.

How is paper wallet commission calculated?

Commission is 0.155% of trade value, plus 15% VAT on the commission — matching standard Saudi brokerage rates. Fees are deducted from your paper cash balance automatically. Tasilab applies no minimum commission; most Saudi brokers do charge one, commonly SAR 12 to SAR 50 per trade, so a real trade below roughly SAR 7,700 in value would cost more than it does here.

How does slippage work on market orders?

Market orders include realistic slippage based on order size vs. average daily volume: 0% under 1% of ADV, 0.1% between 1–5%, and 0.25% above 5%. Limit orders have no slippage and fill at the exact limit price.

Export & share

How do I export an Edge Explorer or Quick Predictor result?

After running a query, the result card has an Export bar with seven options: Copy JSON (clipboard, ready to paste anywhere), PDF (the full research report — the Pro tier, with methodology and a reproduction snippet), PNG · quick (question, verdict and the numbers behind it — built to be read in seconds), PNG · detailed (1080×1350 portrait with the matched-vs-baseline table and a chart), PNG · link preview (1200×630, ideal for messaging apps and social), PNG · square (1080×1080, for Instagram / WhatsApp status), and Share link (next FAQ).

Can I really paste the JSON straight into ChatGPT or Claude?

Yes — that's exactly what it's designed for. The exported JSON is a self-contained Report envelope: a one-paragraph plain-English synopsis at the top, full structured payload below, plus the methodological caveats. Try a prompt like "Here's a Tasilab research report. What's the most likely overfitting risk and how would you stress-test the result?" followed by the JSON. The same envelope can also reach the API at GET /v1/reports/{kind}/{run_id} if you want to pull it programmatically.

How do I share a result on WhatsApp, Telegram, or X?

Click Share link on the export bar. We mint a public URL like https://api.tasilab.com/share/ml/<token>; on phones, your OS share sheet pops directly so you can pick WhatsApp / Telegram / etc. — the message will auto-render with the share-card preview thanks to Open Graph + Twitter Card meta tags. On desktop the link goes to your clipboard. The link is read-only; viewers see the rendered card and the synopsis but cannot see your account, parameters not in the result, or any other run.

How do I revoke a public share link?

Click the Revoke button next to Share link on the export bar. The token is soft-revoked immediately; anyone hitting the URL gets a 404 within ~5 minutes (cache TTL). Re-sharing later mints a fresh token — old revoked URLs stay dead.

Where is my run after I close the page?

Every Edge Explorer / Quick Predictor result is persisted with a stable run_id. You can fetch it again any time at /v1/reports/{kind}/{run_id} (JSON), .pdf, or .png — same content as the export buttons, just from the API directly.

Quick Predictor

One-click classifier on a fixed feature set (lagged returns, RSI, volume ratio, overnight gap). Predicts next-day direction. Reports out-of-sample accuracy + AUC + feature importances. Useful as a sanity check before you build a strategy around an idea — if the predictor can't even beat the majority-class baseline, the features probably aren't carrying signal.

Symbol

History window

Model

Result

Finding

One train/test split on one stock. Held-out accuracy, not a prediction you can trade.

What was tested—

Rows held out—

Window searched—

Against always-guess baseline—

Out-of-sample (test set)

Accuracy—

AUC—

Baseline (majority)—

Test rows—

Training (in-sample)

Accuracy—

Train rows—

Fit time—

Model—

ROC curve (test set)

Confusion matrix (test set)

Feature importance

What this shows

Whether a standard classifier on a fixed feature set can separate up days from down days on rows it was never trained on.

What this does not show

A tradeable edge. Accuracy ignores fees, slippage and position sizing, and this is a single chronological split, not cross-validation — one lucky test period can flatter the score.

Run details & reproduction

Run ID—

Method version—

Data fingerprint—

—

Your experiments — all types

All

Technical indicator

Machine learning

Name	Type	Symbol	Status	Return	Win rate	Trades	Created
Sign in to view your experiments.

Report

1 · Level of detail

Quick

Detailed

Full

2 · Format

PNG

PDF

Preview

Your saved runs

Every Edge Explorer, Quick Predictor and Backtest run you have saved. Export any of them again — the report is rebuilt from the stored run, so it says exactly what it said the day you ran it.

All

Edge Explorer

Quick Predictor

Backtest

Run	Surface	Symbol	Window	Verdict	Created	Export
Sign in to view your saved runs.

By the numbers

Plan

API key

—

Reveal, rotate, or read the docs

Contact

Settings

Language

Reduce motion—

Notifications—

Data sourcesThird-party provider

Support

Not in this versionvNext

Mobile verification by SMS / WhatsApp codevNext

Connect X and TradingViewvNext

Public profilevNext

No backend exists for these yet — so they stay designed, not shipped.

Sign out

This removes your API key from the browser. Your account and trade history stay safe on the server.

Delete account

This immediately and irreversibly removes your email, password, and any linked Google or Apple sign-in. Your trading history and Majlis ledger are kept anonymized — never erased, never traceable back to you.

Tasilab · Simulation-only, not investment advice

Paper-trade the Saudi Exchange.

Sign in to Tasilab

Before you begin

Welcome to Tasilab. Let's run something.

Let's get you set up.

No runs in flight

No completed experiments yet

No trades yet