Interactive Brokers API Tutorial 2026: Connect, Trade, Stream

Why Interactive Brokers?

Interactive Brokers (IBKR) is the dominant retail and small-fund algorithmic trading broker globally. Three reasons:

1. Coverage. Access to 150+ exchanges across 33 countries, plus options, futures, FX, bonds, mutual funds and crypto.
2. Cost. Per-share fees ($0.005/share) are competitive even for high-frequency strategies; for retail-scale algo trading, costs are typically 30-70% lower than at most competitors.
3. API quality. The API has flaws (which we'll cover) but it's mature, well-documented, and supports virtually every order type and asset class IB offers.

This tutorial covers the practical realities of connecting to IB's API in 2026, with Python code examples for the most common workflows. For broader broker comparison, see our best brokers for algo trading.

---

Account Setup

Before writing code, you need:

1. An IB account. Either Individual or Pro. Pro accounts have lower data fees and better API access; Individual accounts work for non-professional traders.
2. TWS (Trader Workstation) or IB Gateway. Both bridge between your code and IB's servers. Gateway is lighter (no GUI) and preferred for algo trading.
3. API enabled in your account settings. In TWS/Gateway: Edit > Global Configuration > API > Settings. Tick "Enable ActiveX and Socket Clients."
4. Set the API port. Default is 7497 for paper trading TWS, 7496 for live TWS, 4002 for paper Gateway, 4001 for live Gateway. Note these for your connection code.

For testing, always start with paper trading. IB provides full paper accounts that mirror live functionality without risking real money.

---

Two Python Libraries

You have two main choices for Python:

ib_insync (recommended for most users)

A friendly Python wrapper around the underlying API. Maintained by Ewald de Wit. Significantly easier to use than the native API.

pip install ib_insync

ibapi (the native API)

IB's official Python library. More verbose; closer to how the underlying API actually works. Use if you need very fine control or are integrating with existing IB-native code.

pip install ibapi
# Or download from IB and install manually for the most current version

This tutorial uses ib_insync for clarity. The patterns translate to ibapi with more boilerplate.

---

Step 1: Connect

from ib_insync import IB

ib = IB()
# Paper trading TWS default port
ib.connect('127.0.0.1', 7497, clientId=1)

print(ib.isConnected())  # True if successful

Notes:

• clientId is a unique integer per connection. If you connect from multiple scripts, each needs a different ID.
• If TWS/Gateway isn't running, connection fails immediately.
• The API connection is over TCP localhost; nothing leaves your machine until TWS/Gateway sends it to IB.

---

Step 2: Get Account Information

account_summary = ib.accountSummary()
for tag in account_summary:
    if tag.tag == 'NetLiquidation':
        print(f"Account value: ${tag.value}")

positions = ib.positions()
for p in positions:
    print(f"{p.contract.symbol}: {p.position} shares")

accountSummary() returns dozens of fields (cash, margin, P&L, etc.). positions() returns current holdings.

---

Step 3: Define a Contract

A "contract" in IB terminology is the financial instrument you want to trade.

US stock

from ib_insync import Stock

apple = Stock('AAPL', 'SMART', 'USD')
ib.qualifyContracts(apple)

SMART is IB's smart routing exchange. qualifyContracts confirms with IB that the contract is valid.

Option

from ib_insync import Option

apple_call = Option('AAPL', '20260117', 200, 'C', 'SMART')
ib.qualifyContracts(apple_call)

'20260117' is the expiry (YYYYMMDD), 200 is the strike, 'C' for call.

Future

from ib_insync import Future

es = Future('ES', '20260321', 'CME')  # March 2026 S&P futures
ib.qualifyContracts(es)

FX

from ib_insync import Forex

eurusd = Forex('EURUSD')
ib.qualifyContracts(eurusd)

---

Step 4: Get Market Data

Snapshot quote

ticker = ib.reqMktData(apple)
ib.sleep(2)  # Wait for data
print(f"Bid: {ticker.bid}, Ask: {ticker.ask}, Last: {ticker.last}")
ib.cancelMktData(apple)

Streaming quotes

ticker = ib.reqMktData(apple)
def on_update(t):
    print(f"{t.contract.symbol}: bid={t.bid}, ask={t.ask}")
ticker.updateEvent += on_update

# Stream for 30 seconds
ib.sleep(30)

ticker.updateEvent -= on_update
ib.cancelMktData(apple)

Historical bars

bars = ib.reqHistoricalData(
    apple,
    endDateTime='',
    durationStr='30 D',
    barSizeSetting='1 day',
    whatToShow='TRADES',
    useRTH=True
)
import pandas as pd
df = pd.DataFrame(bars)
print(df.head())

whatToShow options: TRADES, BID, ASK, MIDPOINT, etc. useRTH=True restricts to regular trading hours.

---

Step 5: Place an Order

Market order

from ib_insync import MarketOrder

order = MarketOrder('BUY', 100)
trade = ib.placeOrder(apple, order)

# Wait for fill
ib.sleep(5)
print(f"Status: {trade.orderStatus.status}")
print(f"Filled: {trade.orderStatus.filled}")
print(f"Avg fill price: {trade.orderStatus.avgFillPrice}")

Limit order

from ib_insync import LimitOrder

order = LimitOrder('BUY', 100, 195.50)
trade = ib.placeOrder(apple, order)

Bracket order (parent + take-profit + stop-loss)

parent = LimitOrder('BUY', 100, 195.50)
parent.transmit = False  # Don't send yet

take_profit = LimitOrder('SELL', 100, 205.00)
take_profit.parentId = parent.orderId
take_profit.transmit = False

stop_loss = StopOrder('SELL', 100, 190.00)
stop_loss.parentId = parent.orderId
stop_loss.transmit = True  # Send all three when this is placed

ib.placeOrder(apple, parent)
ib.placeOrder(apple, take_profit)
ib.placeOrder(apple, stop_loss)

Cancel an order

ib.cancelOrder(trade.order)

---

Step 6: Handle Fills and Errors

Subscribe to fills

def on_fill(trade, fill):
    print(f"Filled: {fill.execution.shares} @ {fill.execution.price}")

ib.fillEvent += on_fill

Subscribe to errors

def on_error(reqId, errorCode, errorString, contract):
    print(f"ERROR {errorCode}: {errorString}")

ib.errorEvent += on_error

Common error codes to watch for:

• 104 - Cannot modify a filled order. Common when retrying.
• 201 - Order rejected. Read the message - usually margin or risk-limit related.
• 2103 - Market data farm connection issue. Usually self-resolves.
• 10089 - Requested market data is not subscribed. You need to subscribe to that data feed.

---

Step 7: Common Patterns

Wait for connection

def wait_for_connection(ib, timeout=30):
    start = time.time()
    while not ib.isConnected() and time.time() - start < timeout:
        ib.sleep(0.5)
    if not ib.isConnected():
        raise TimeoutError("Could not connect to IB")

Reconnect on disconnect

def on_disconnect():
    print("Disconnected. Reconnecting...")
    ib.connect('127.0.0.1', 7497, clientId=1)

ib.disconnectedEvent += on_disconnect

Place an order with retry

def safe_place(contract, order, retries=3):
    for attempt in range(retries):
        try:
            trade = ib.placeOrder(contract, order)
            ib.sleep(0.5)
            if trade.orderStatus.status not in ('Cancelled', 'ApiCancelled'):
                return trade
        except Exception as e:
            print(f"Attempt {attempt+1} failed: {e}")
            ib.sleep(1)
    raise RuntimeError(f"Could not place order after {retries} retries")

---

Common Pitfalls

1. Pacing violations

IB throttles API requests. Common limits:

• 50 messages per second
• 60 historical data requests per 10 minutes
• 100 messages per second total

Hitting these triggers errors. Add throttling for high-volume use cases:

from time import sleep
for symbol in symbols:
    bars = ib.reqHistoricalData(...)
    sleep(0.2)  # Stay under request rate

2. Market data subscriptions

IB requires you to subscribe (and pay) for market data feeds. Most casual algo traders only need US Equities (free for $30+/month account value). Options data adds cost; international markets add cost.

3. Threading

ib_insync is asyncio-based. Don't mix with threading without care. If you need parallel work, use asyncio.gather() or separate IB instances with different clientIds.

4. Time zones

IB times are UTC by default. Convert carefully:

from zoneinfo import ZoneInfo
ny = ZoneInfo('America/New_York')
local_time = bar.date.astimezone(ny)

5. Don't run live code accidentally

Always check: connect to port 7497 (paper) when developing, not 7496 (live). One bad clientId mix-up can place real orders.

---

Where to Go From Here

For complete algo trading workflows beyond just the IB API:

• Backtesting platforms comparison - choose a platform that integrates with IB
• QuantConnect review - QuantConnect uses IB as its primary live broker
• Best brokers for algo trading - alternatives if IB doesn't fit
• Pairs trading tutorial - example strategy implementation
• Statistical arbitrage strategies - more complex strategy patterns

For broader Python skills:

• Python for finance guide
• Python fundamentals for quant finance