UNICORN Binance WebSocket API

Description | Installation | Documentation | Examples | Change Log | Wiki | Social | Notifications | Bugs | Contributing | Disclaimer

A Python SDK to use the Binance Websocket API`s (com+testnet, com-margin+testnet, com-isolated_margin+testnet, com-futures+testnet, com-coin_futures, com-vanilla-options+testnet, com-portfolio_margin, us, tr) in a simple, fast, flexible, robust and fully-featured way.

Part of 'UNICORN Binance Suite'.

Receive Data from Binance WebSockets

Create a multiplex websocket connection to Binance with a stream_buffer with just 3 lines of code

from unicorn_binance_websocket_api import BinanceWebSocketApiManager

ubwa = BinanceWebSocketApiManager(exchange="binance.com")
ubwa.create_stream(channels=['trade', 'kline_1m'], markets=['btcusdt', 'bnbbtc', 'ethbtc'])

And 4 more lines to print out the data

while True:
    oldest_data_from_stream_buffer = ubwa.pop_stream_data_from_stream_buffer()
    if oldest_data_from_stream_buffer:
        print(oldest_data_from_stream_buffer)

Or with a callback function just do

from unicorn_binance_websocket_api import BinanceWebSocketApiManager

def process_new_receives(stream_data):
    print(str(stream_data))

ubwa = BinanceWebSocketApiManager(exchange="binance.com")
ubwa.create_stream(channels=['trade', 'kline_1m'], 
                   markets=['btcusdt', 'bnbbtc', 'ethbtc'], 
                   process_stream_data=process_new_receives)

Or with an async callback function just do

from unicorn_binance_websocket_api import BinanceWebSocketApiManager
import asyncio

async def process_new_receives(stream_data):
    print(stream_data)
    await asyncio.sleep(1)

ubwa = BinanceWebSocketApiManager()
ubwa.create_stream(channels=['trade', 'kline_1m'],
                   markets=['btcusdt', 'bnbbtc', 'ethbtc'],
                   process_stream_data_async=process_new_receives)

Or await the stream data in an asyncio coroutine

All the methods of data collection presented have their own advantages and disadvantages. However, this is the generally recommended method for processing data from streams.

from unicorn_binance_websocket_api import BinanceWebSocketApiManager
import asyncio

async def main():
    async def process_asyncio_queue(stream_id=None):
        print(f"Start processing the data from stream '{ubwa.get_stream_label(stream_id)}':")
        while ubwa.is_stop_request(stream_id) is False:
            data = await ubwa.get_stream_data_from_asyncio_queue(stream_id)
            print(data)
            ubwa.asyncio_queue_task_done(stream_id)
    ubwa.create_stream(channels=['trade'],
                       markets=['ethbtc', 'btcusdt'],
                       stream_label="TRADES",
                       process_asyncio_queue=process_asyncio_queue)
    while not ubwa.is_manager_stopping():
        await asyncio.sleep(1)

with BinanceWebSocketApiManager(exchange='binance.com') as ubwa:
    try:
        asyncio.run(main())
    except KeyboardInterrupt:
        print("\r\nGracefully stopping ...")
    except Exception as e:
        print(f"\r\nERROR: {e}\r\nGracefully stopping ...")

Basically that's it, but there are more options.

Receive private UserData Streams

Create a private !userData stream to receive account updates like order fills, balance changes and position updates in real time. A valid api_key and api_secret is required.

Set the credentials globally on the manager

All streams created on this manager inherit the credentials:

from unicorn_binance_websocket_api import BinanceWebSocketApiManager

async def process_userdata(stream_data):
    print(stream_data)

ubwa = BinanceWebSocketApiManager(exchange="binance.com",
                                  api_key="YOUR_BINANCE_API_KEY",
                                  api_secret="YOUR_BINANCE_API_SECRET")
ubwa.create_stream(channels='arr',
                   markets='!userData',
                   process_stream_data_async=process_userdata)

Or pass the credentials per stream

Useful when running multiple !userData streams with different API keys on the same manager:

ubwa = BinanceWebSocketApiManager(exchange="binance.com")

ubwa.create_stream(channels='arr',
                   markets='!userData',
                   api_key="API_KEY_ACCOUNT_A",
                   api_secret="API_SECRET_ACCOUNT_A",
                   stream_label="ACCOUNT_A",
                   process_stream_data_async=process_userdata)

ubwa.create_stream(channels='arr',
                   markets='!userData',
                   api_key="API_KEY_ACCOUNT_B",
                   api_secret="API_SECRET_ACCOUNT_B",
                   stream_label="ACCOUNT_B",
                   process_stream_data_async=process_userdata)

Per-stream credentials override the manager defaults. Isolated Margin additionally requires the symbols parameter:

ubwa_im = BinanceWebSocketApiManager(exchange="binance.com-isolated_margin")
ubwa_im.create_stream(channels='arr',
                      markets='!userData',
                      symbols='btcusdt',
                      api_key="YOUR_BINANCE_API_KEY",
                      api_secret="YOUR_BINANCE_API_SECRET",
                      process_stream_data_async=process_userdata)

See also example_multiple_userdata_streams.py.

Convert received stream data into well-formed Python dictionaries with UnicornFy

unicorn_fied_stream_data = UnicornFy.binance_com_websocket(data)

or

ubwa.create_stream(['trade'], ['btcusdt'], output="UnicornFy")

Subscribe / unsubscribe new markets and channels

markets = ['engbtc', 'zileth']
channels = ['kline_5m', 'kline_15m', 'kline_30m', 'kline_1h', 'kline_12h', 'depth5']

ubwa.subscribe_to_stream(stream_id=stream_id, channels=channels, markets=markets)

ubwa.unsubscribe_from_stream(stream_id=stream_id, markets=markets)

ubwa.unsubscribe_from_stream(stream_id=stream_id, channels=channels)

Send Requests to Binance WebSocket API

Place orders, cancel orders or send other requests via WebSocket

from unicorn_binance_websocket_api import BinanceWebSocketApiManager

api_key = "YOUR_BINANCE_API_KEY"
api_secret = "YOUR_BINANCE_API_SECRET"

async def process_api_responses(stream_id=None):
    while ubwa.is_stop_request(stream_id=stream_id) is False:
        data = await ubwa.get_stream_data_from_asyncio_queue(stream_id=stream_id)
        print(data)
        ubwa.asyncio_queue_task_done(stream_id=stream_id)

ubwa = BinanceWebSocketApiManager(exchange="binance.com")
api_stream = ubwa.create_stream(api=True,
                                api_key=api_key,
                                api_secret=api_secret,
                                output="UnicornFy",
                                process_asyncio_queue=process_api_responses)

response = ubwa.api.spot.get_server_time(return_response=True)
print(f"Binance serverTime: {response['result']['serverTime']}")

orig_client_order_id = ubwa.api.spot.create_order(order_type="LIMIT",
                                                  price = 1.1,
                                                  quantity = 15.0,
                                                  side = "SELL",
                                                  symbol = "BUSDUSDT")

ubwa.api.spot.cancel_order(orig_client_order_id=orig_client_order_id, symbol="BUSDUSDT")                                   

All available methods:

• Futures
• Spot

Here you can find a complete guide on how to process requests via the Binance WebSocket API!

Stop ubwa after usage to avoid memory leaks

When you instantiate UBWA with with, ubwa.stop_manager() is automatically executed upon exiting the with-block.

with BinanceWebSocketApiManager() as ubwa:
    ubwa.create_stream(channels="trade", markets="btcusdt", stream_label="TRADES")

Without with, you must explicitly execute ubwa.stop_manager() yourself.

ubwa.stop_manager()

stream_signals - know the state of your streams

Usually you want to know when a stream is working and when it is not. This can be useful to know that your own system is currently "blind" and you may want to close open positions to be on the safe side, know that indicators will now provide incorrect values or that you have to reload the missing data via REST as an alternative.

For this purpose, the UNICORN Binance WebSocket API provides so-called stream_signals, which are used to tell your code in real time when a stream is connected, when it received its first data record, when it was disconnected and stopped, and when the stream cannot be restored.

from unicorn_binance_websocket_api import BinanceWebSocketApiManager
import time

def process_stream_signals(signal_type=None, stream_id=None, data_record=None, error_msg=None):
    print(f"Received stream_signal for stream '{ubwa.get_stream_label(stream_id=stream_id)}': "
          f"{signal_type} - {stream_id} - {data_record} - {error_msg}")

with BinanceWebSocketApiManager(process_stream_signals=process_stream_signals) as ubwa:
    ubwa.create_stream(channels="trade", markets="btcusdt", stream_label="TRADES")
    print(f"Waiting a few seconds and then stopping the stream ...")
    time.sleep(7)

More?

Discover even more possibilities, use this script to stream everything from "binance.com" or try our examples!

This should be known by everyone using this lib:

• Best practice solutions for a maximum stable connection
• Do you want consistent data from binance?

Description

The Python package UNICORN Binance WebSocket API provides an API to the Binance Websocket API`s of Binance (+Testnet), Binance Margin (+Testnet), Binance Isolated Margin (+Testnet), Binance Futures (+Testnet), Binance COIN-M Futures, Binance European Options (+Testnet), Binance US and Binance TR and supports sending requests to the Binance Websocket API and the streaming of all public streams like trade, kline, ticker, depth, bookTicker, forceOrder, compositeIndex etc. and also all private userData streams which needs to be used with a valid api_key and api_secret from the Binance Exchange www.binance.com, testnet.binance.vision or www.binance.us.

Use the UNICORN Binance REST API in combination.

What are the benefits of the UNICORN Binance WebSocket API?

• Fully managed websockets and 100% auto-reconnect! Also handles maintenance windows!

• No memory leaks from Python version 3.9 to 3.14!

• The full UBS stack is delivered as a compiled C extension for maximum performance.

• Support for Binance Websocket API, send requests like create_order, cancel_open_orders and many more directly over websocket!

• Supported exchanges:

Exchange	Exchange string	WS	WS API
Binance	binance.com
Binance Testnet	binance.com-testnet
Binance Margin	binance.com-margin
Binance Margin Testnet	binance.com-margin-testnet
Binance Isolated Margin	binance.com-isolated_margin
Binance Isolated Margin Testnet	binance.com-isolated_margin-testnet
Binance USD-M Futures	binance.com-futures
Binance USD-M Futures Testnet	binance.com-futures-testnet
Binance Coin-M Futures	binance.com-coin_futures
Binance European Options	binance.com-vanilla-options
Binance European Options Testnet	binance.com-vanilla-options-testnet
Binance Portfolio Margin*	binance.com-portfolio_margin
Binance US	binance.us
Binance TR	trbinance.com

* Portfolio Margin support is currently limited to the user data stream (!userData), see the Portfolio Margin example and issue #452.

• Streams are processing asynchronous/concurrent (Python asyncio) and each stream is started in a separate thread, so you don't need to deal with asyncio in your code! But you can consume with await , if you want!

• Supports subscribe/unsubscribe on all exchanges! (Take a look to the max supported subscriptions per stream in the endpoint configuration overview!)

• UNICORN Binance WebSocket API respects Binance's API guidelines and protects you from avoidable reconnects and bans.

• Support for multiple private !userData streams with different api_key and api_secret. (example_multiple_userdata_streams.py)

• Pick up the received data from the stream_buffer (FIFO or LIFO) - if you can not store your data in cause of a temporary technical issue, you can kick back the data to the stream_buffer which stores the receives in the RAM till you are able to process the data in the normal way again. Learn more!

• Use separate stream_buffers for specific streams or users!

• Watch the stream_signals to receive CONNECT, FIRST_RECEIVED_DATA, DISCONNECT, STOP and STREAM_UNREPAIRABLE signals from your streams! Learn more!

• Get the received data unchanged as received, as Python dictionary or converted with UnicornFy into well-formed Python dictionaries. Use the output parameter of create_stream() to control the output format.

• Helpful management features like clear_asyncio_queue(), clear_stream_buffer(), get_binance_api_status(), get_current_receiving_speed(), get_errors_from_endpoints(), get_limit_of_subscriptions_per_stream(), get_request_id(), get_result_by_request_id(), get_results_from_endpoints(), get_stream_buffer_length(), get_stream_info(), get_stream_list(), get_stream_id_by_label(), get_stream_statistic(), get_stream_subscriptions(), get_version(), is_update_available(), get_stream_data_from_asyncio_queue(), pop_stream_data_from_stream_buffer(), print_summary(), replace_stream(), set_stream_label(), set_ringbuffer_error_max_size(), subscribe_to_stream(), stop_stream(), unsubscribe_from_stream(), wait_till_stream_has_started() and many more! Explore them here.

• Monitor the status of the created BinanceWebSocketApiManager() instance within your code with get_monitoring_status_plain() and specific streams with get_stream_info().

• Available as a package via pip and conda as precompiled C extension with stub files for improved Intellisense functions and source code for easier debugging of the source code. To the installation.

• Nice to use with iPython: "IPython (Interactive Python) is a command shell for interactive computing that offers introspection, rich media, shell syntax, tab completion, and history." (example_interactive_mode.py)

• Also, nice to use with the Jupyter Notebook :)

• Integration of test cases and examples.

• Customizable base URL.

• Choice of the WebSocket engine: websockets (default) or picows, see WebSocket library.

• Proxy support (HTTP, HTTPS, SOCKS4, SOCKS5), passed natively to the WebSocket library:

  ubwa = BinanceWebSocketApiManager(exchange="binance.com", proxy="socks5://user:pass@127.0.0.1:9050")
  ubwa = BinanceWebSocketApiManager(exchange="binance.com", proxy="http://127.0.0.1:3128")
  

  The legacy socks5_proxy_server/socks5_proxy_user/socks5_proxy_pass parameters keep working. REST requests (listenKey handling) follow a socks5:// proxy only. Credentials containing @, :, / or % need percent-encoding in the URL; websockets sends them without decoding (python-websockets/websockets#1761), so UBWA rejects such credentials for websocket_library="websockets" at construction - use plain credentials or picows.

  Read the docs or this how to for more information or try example_socks5_proxy.py.

• Excessively tested on Linux, Mac and Windows on x86, arm32, arm64, ...

If you like the project, please it on GitHub!

Installation and Upgrade

The module requires Python 3.9 and runs smoothly up to and including Python 3.14.

PyPy wheels are available for all supported Python versions.

conda-forge note: Conda packages are provided for Python 3.10 – 3.14. Python 3.9 is not available on conda-forge — it was dropped from the global pinning after reaching end-of-life in October 2025. For Python 3.9, use pip install.

The current dependencies are listed here.

If you run into errors during the installation take a look here.

Packages are created automatically with GitHub Actions

When a new release is created, the Build and Publish GH+PyPi workflow spins up virtual Windows/Linux/Mac runners, compiles the Cython extensions, builds the wheels and publishes them on GitHub and PyPI. The conda-forge feedstock conda-forge/unicorn-binance-websocket-api-feedstock picks up the new PyPI release automatically and builds the Conda packages on its own infrastructure. This is a transparent method that makes it possible to trace the source code behind a compilation.

A Cython binary, PyPy or source code based CPython wheel of the latest version with pip from PyPI

Our Cython and PyPy Wheels are available on PyPI, these wheels offer significant advantages for Python developers:

• Performance Boost with Cython Wheels: Cython is a programming language that supplements Python with static typing and C-level performance. By compiling Python code into C, Cython Wheels can significantly enhance the execution speed of Python code, especially in computationally intensive tasks. This means faster runtimes and more efficient processing for users of our package.

• PyPy Wheels for Enhanced Efficiency: PyPy is an alternative Python interpreter known for its speed and efficiency. It uses Just-In-Time (JIT) compilation, which can dramatically improve the performance of Python code. Our PyPy Wheels are tailored for compatibility with PyPy, allowing users to leverage this speed advantage seamlessly.

Both Cython and PyPy Wheels on PyPI make the installation process simpler and more straightforward. They ensure that you get the optimized version of our package with minimal setup, allowing you to focus on development rather than configuration.

On Raspberry Pi and other architectures for which there are no pre-compiled versions, the package can still be installed with PIP. PIP then compiles the package locally on the target system during installation. Please be patient, this may take some time!

Installation

pip install unicorn-binance-websocket-api

Update

pip install unicorn-binance-websocket-api --upgrade

conda

conda install -c conda-forge unicorn-binance-websocket-api

WebSocket library: websockets or picows

UBWA uses the websockets library by default. Alternatively it can run on picows, a Cython implementation of the WebSocket protocol that ships a drop-in replacement of the websockets client API (picows.websockets). picows is an optional dependency:

pip install unicorn-binance-websocket-api[picows]

Select the library per manager instance, everything else stays the same:

ubwa = BinanceWebSocketApiManager(exchange="binance.com", websocket_library="picows")

Selecting "picows" without the package installed raises an ImportError, an unknown value raises a ValueError - there is no silent fallback. Proxies (proxy="http://...", https://, socks4://, socks5://) are passed to both libraries natively. The conda-forge package is picows.

picows support is new and opt-in: websockets stays the default until picows has proven itself in real-world use and enough reports are in (the first upstream finding, tarasko/picows#108, is fixed in picows 2.2.0; the extra requires picows 2.3.0 for its native proxy support). Questions, experiences and your own benchmark numbers: issue #477 - WebSocket library: websockets vs. picows. Background, benchmarks and the 24 h soak in the article picows in UNICORN Binance WebSocket API: Up to 2× the Throughput, Opt-In for Now.

Is picows faster? Measured, not assumed

dev/test_websocket_library_benchmark.py replays Binance shaped messages from a local server (separate process) through the complete UBWA stack (connection → stream loop → process_stream_data callback), 3 runs, median. Python 3.13, websockets 16.0, picows 2.3.0, x86_64 Linux (8 cores), output_default="raw_data":

• Up to ~1 KB per message (aggTrade, kline, bookTicker, depth20, ...) picows delivers 1.7x-2x the throughput at about half the CPU per message. At 9 KB (full depth diffs) picows is 1.2x ahead, and on the 450 KB !ticker@arr payload it is behind. That row is an artifact of the local replay (a loopback firehose feeding a consumer that is slower than the wire, so picows drains a multi-MB socket buffer in one read): driven directly picows wins at every size, and with --rcvbuf 131072 (client socket receive buffer capped, closer to a WAN link) it is 1.4x-1.55x ahead through UBWA as well. Details in context/websocket-library.md.
• With output_default="dict" (orjson parsing included) the gap is 1.4x-1.7x for messages up to 1 KB.
• Against live binance.com with a 20 symbol multiplex (a few hundred msgs/s) the choice makes no measurable difference: the CPU load is dominated by UBWA's fixed per-manager overhead, not by the transport.
• So: pick picows for high-throughput consumers (many streams, depth@100ms on hundreds of symbols, CPU-bound hosts), stay on websockets if you need the default, its broader ecosystem or PyPy. Full tables including the raw-library baseline and the live run: context/websocket-library.md; what UBWA itself costs per message and how that was cut: context/stream-loop.md.

From source of the latest release with PIP from GitHub

Linux, macOS, ...

Run in bash:

pip install https://github.com/oliver-zehentleitner/unicorn-binance-websocket-api/archive/$(curl -s https://api.github.com/repos/oliver-zehentleitner/unicorn-binance-websocket-api/releases/latest | grep -oP '"tag_name": "\K(.*)(?=")').tar.gz --upgrade

Windows

Use the below command with the version (such as 2.16.1) you determined here:

pip install https://github.com/oliver-zehentleitner/unicorn-binance-websocket-api/archive/2.16.1.tar.gz --upgrade

From the latest source (dev-stage) with PIP from GitHub

This is not a release version and can not be considered to be stable!

pip install https://github.com/oliver-zehentleitner/unicorn-binance-websocket-api/tarball/master --upgrade

Change Log

https://oliver-zehentleitner.github.io/unicorn-binance-websocket-api/changelog.html

Documentation

• General
• Modules

Examples

• Look here!

Related Articles

• The Complete Binance Python API Guide 2026
• How to create a Binance API Key and API Secret?
• Create and Cancel Orders via WebSocket on Binance
• How to Download Klines from Binance using Python?
• Passing Binance Market Data to Apache Kafka in Python with aiokafka
• How to Connect to binance.com Websockets using Python via a Socks5 Proxy
• When IP Whitelisting Isn't What It Seems: A Real-World Case Study from the Binance API
• UBDCC Deep Dive: Building a Trust Layer for Binance Order Books
• picows in UNICORN Binance WebSocket API: Up to 2× the Throughput, Opt-In for Now
• UNICORN Binance Suite Article Series

Project Homepage

https://github.com/oliver-zehentleitner/unicorn-binance-websocket-api

Wiki

https://github.com/oliver-zehentleitner/unicorn-binance-websocket-api/wiki

Social

• Discussions
• Telegram
• Reddit Community
• https://dev.binance.vision

Receive Notifications

To receive notifications on available updates you can the repository on GitHub, write your own script with using is_update_available() or you use get_monitoring_status_plain().

To receive news (like inspection windows/maintenance) about the Binance API`s subscribe to their telegram groups:

• https://t.me/binance_api_announcements
• https://t.me/binance_api_english
• https://t.me/Binance_USA
• https://t.me/TRBinanceTR
• https://t.me/BinanceExchange

How to report Bugs or suggest Improvements?

List of planned features - click if you need one of them or suggest a new feature!

Before you report a bug, try the latest release. If the issue still exists, provide the error trace, OS and Python version and explain how to reproduce the error. A demo script is appreciated.

If you don't find an issue related to your topic, please open a new issue!

Report a security bug!

Contributing

UNICORN Binance WebSocket API is an open source project which welcomes contributions which can be anything from simple documentation fixes and reporting dead links to new features. To contribute follow this guide.

Contributors

We open source!

AI Integration

This project provides a llms.txt file for AI tools (ChatGPT, Claude, Copilot, etc.) with structured usage instructions, code examples and module routing.

Disclaimer

This project is for informational purposes only. You should not construe this information or any other material as legal, tax, investment, financial or other advice. Nothing contained herein constitutes a solicitation, recommendation, endorsement or offer by us or any third party provider to buy or sell any securities or other financial instruments in this or any other jurisdiction in which such solicitation or offer would be unlawful under the securities laws of such jurisdiction.

If you intend to use real money, use it at your own risk!

Under no circumstances will we be responsible or liable for any claims, damages, losses, expenses, costs or liabilities of any kind, including but not limited to direct or indirect damages for loss of profits.

SOCKS5 Proxy / Geoblocking

We would like to explicitly point out that in our opinion US citizens are exclusively authorized to trade on Binance.US and that this restriction must not be circumvented!

The purpose of supporting a SOCKS5 proxy in the UNICORN Binance Suite and its modules is to allow non-US citizens to use US services. For example, GitHub actions with UBS will not work without a SOCKS5 proxy, as they will inevitably run on servers in the US and be blocked by Binance.com. Moreover, it also seems justified that traders, data scientists and companies from the US analyze binance.com market data - as long as they do not trade there.