Getting Started#


Alpaca-py provides an interface for interacting with the API products Alpaca offers. These API products are provided as various REST, WebSocket and SSE endpoints that allow you to do everything from streaming market data to creating your own trading apps. You can learn about the API products Alpaca offers at


Alpaca’s APIs allow you to do everything from building algorithmic trading strategies to building a full brokerage experience for your own end users. Here are some things you can do with Alpaca-py.

Market Data API: Access live and historical market data for 5000+ stocks and 20+ crypto.

Trading API: Trade stock and crypto with lightning fast execution speeds.

Broker API & Connect: Build investment apps - from robo-advisors to brokerages.


Alpaca-py is supported on Python 3.7+. You can install Alpaca-py using pip. To learn more about version histories, visit the PyPI page.

To install Alpaca-py, run the following pip command in your terminal.

pip install alpaca-py


Try upgrading your pip before installing if you face errors.

pip install --upgrade pip


If you’re using poetry to manage dependencies in your project. You can add Alpaca-py to your project by running

poetry add alpaca-py

What’s New?#

If you’ve used the previous python SDK alpaca-trade-api, there are a few key differences to be aware of.

Broker API#

Alpaca-py lets you use Broker API to start building your investment apps! Learn more at the Broker page.

OOP Design#

Alpaca-py uses a more OOP approach to submitting requests compared to the previous SDK. To submit a request, you will most likely need to create a request object containing the desired request data. Generally, there is a unique request model for each method.

Some examples of request models corresponding to methods:

  • GetOrdersRequest for TradingClient.get_orders()

  • CryptoLatestOrderbookRequest for CryptoHistoricalDataClient.get_crypto_latest_orderbook()

Request Models Usage Example

To get historical bar data for crypto, you will need to provide a CryptoBarsRequest object.

from import CryptoHistoricalDataClient
from import CryptoBarsRequest
from import TimeFrame

# no keys required for crypto data
client = CryptoHistoricalDataClient()

request_params = CryptoBarsRequest(
                        symbol_or_symbols=["BTC/USD", "ETH/USD"],

bars = client.get_crypto_bars(request_params)

Data Validation#

Alpaca-py uses pydantic to validate data models at run-time. This means if you are receiving request data via JSON from a client. You can handle parsing and validation through Alpaca’s request models. All request models can be instantiated by passing in data in dictionary format.

Here is a rough example of what is possible.

@app.route('/post_json', methods=['POST'])
def do_trade():
    # ...

    order_data_json = request.get_json()

    # validate data

    # ...

Many Clients#

Alpaca-py has a lot of client classes. There is a client for each API and even asset class specific clients (StockHistoricalData, CryptoDataStream). This requires you to pick and choose clients based on your needs.

Broker API: BrokerClient

Trading API: TradingClient

Market Data API: StockHistoricalDataClient, CryptoHistoricalDataClient, CryptoDataStream, StockDataStream

API Keys#

Trading and Market Data API#

In order to use Alpaca’s services you’ll need to sign up for an Alpaca account and retrieve your API keys. Signing up is completely free and takes only a few minutes. Sandbox environments are available to test out the API. To use the sandbox environment, you will need to provide sandbox/paper keys. API keys are passed into Alpaca-py through either TradingClient, StockHistoricalDataClient, CryptoHistoricalDataClient, StockDataStream, or CryptoDataStream.

Broker API#

To use the Broker API, you will need to sign up for a broker account and retrieve your Broker API keys. The API keys can be found on the dashboard once you’ve logged in. Alpaca also provides a sandbox environment to test out Broker API. To use the sandbox mode, provide your sandbox keys. Once you have your keys, you can pass them into BrokerClient to get started.