Due to an increase in demand, you may experience delays with Live Support. If you're having trouble signing in, please view this guide.
Example code for interpreting the WebSocket API v1 order book feed (asynchronous/multi-threaded)
Detailed instructions for interpreting the WebSocket API v1 order book feed are available in our How to maintain a valid order book support page.
Example code in Python for integrating with the WebSocket API v1 order book feed and maintaining a local copy of an order book is provided below.
The code uses the standard Python WebSocket library and implements an asynchronous (event driven/multi-threaded) approach, which allows other tasks to continue in parallel to the order book feed interpretation. 


The code should be executed via a command prompt (such as macOS Terminal) as follows:
  • krakenwsorderbookasync.py symbol depth
  • or
  • python3 krakenwsorderbookasync.py symbol depth
For example, the following command would retrieve and interpret the Dogecoin/USD (XDG/USD) order book feed at a depth of 10 price levels (per side):
  • krakenwsorderbookasync.py XDG/USD 10
Note that the code will run indefinitely until terminated by a keyboard interrupt (Ctrl+C in most command line environments).

Python code

The Python code can be viewed below and can also be downloaded as krakenwsorderbookasync.py.
#!/usr/bin/env python3

# Import WebSocket client library (and others)
import sys
import json
import signal
import time
import _thread
import websocket

# Parse command line arguments (symbol and depth)
if len(sys.argv) < 3:
    api_symbol = sys.argv[1]
    api_depth = int(sys.argv[2])

# Define order book variables
api_book = {'bid':{}, 'ask':{}}

# Define order book update functions
def dicttofloat(data):
    return float(data[0])

def api_book_update(api_book_side, api_book_data):
    for data in api_book_data:
        price_level = data[0]
        volume = data[1]
        if float(volume) > 0.0:
            api_book[api_book_side][price_level] = volume
        if api_book_side == 'bid':
            api_book['bid'] = dict(sorted(api_book['bid'].items(), key=dicttofloat, reverse=True)[:api_depth])
        elif api_book_side == 'ask':
            api_book['ask'] = dict(sorted(api_book['ask'].items(), key=dicttofloat)[:api_depth])

# Define WebSocket callback functions
def ws_thread(*args):
    ws = websocket.WebSocketApp('wss://ws.kraken.com/', on_open=ws_open, on_message=ws_message)

def ws_open(ws):
    ws.send('{"event":"subscribe", "subscription":{"name":"book", "depth":%(api_depth)d}, "pair":["%(api_symbol)s"]}' % {'api_depth':api_depth, 'api_symbol':api_symbol})

def ws_message(ws, ws_data):
    api_data = json.loads(ws_data)
    if 'event' in api_data:
        if 'as' in api_data[1]:
            api_book_update('ask', api_data[1]['as'])
            api_book_update('bid', api_data[1]['bs'])
            for data in api_data[1:len(api_data)-2]:
                if 'a' in data:
                    api_book_update('ask', data['a'])
                elif 'b' in data:
                    api_book_update('bid', data['b'])

# Start new thread for WebSocket interface
_thread.start_new_thread(ws_thread, ())

# Output order book (once per second) in main thread
    while True:
        if len(api_book['bid']) < api_depth or len(api_book['ask']) < api_depth:
            bid = sorted(api_book['bid'].items(), key=dicttofloat, reverse=True)
            ask = sorted(api_book['ask'].items(), key=dicttofloat)
            for count in range(api_depth):
                print('%(bidprice)s (%(bidvolume)s)\t\t\t\t%(askprice)s (%(askvolume)s)' \
                      % {'bidprice':bid[count][0], 'bidvolume':bid[count][1], 'askprice':ask[count][0], 'askvolume':ask[count][1]})
except KeyboardInterrupt: