How to Make a Chat Application in Python

Learn how to build a simple chat room application that accepts multiple connected clients using built-in's sockets and threading libraries in Python.
  · 8 min read · Updated oct 2020 · Python Standard Library

A chat room is an interface that allows two or more people to chat and send messages to everyone in the room. In this tutorial, you will learn how to build a simple chat room server and allow multiple clients to connect to it using sockets in Python.

We are going to use socket module which comes built-in with Python and provides us with socket operations that are widely used on the Internet, as they are behind of any connection to any network.

To get started, and for fancy printing, we gonna need colorama package to assign a printing color to each client in the chatroom:

pip3 install colorama

Since we're using sockets, then we need a server and client code, let's start with the server side.

Server Code

In our architecture, the whole job of the server is to do two essential operations:

  • Listening for upcoming client connections, if a new client is connected, we add it to our collection of client sockets.
  • Start a new thread for each client connected that keep listening for upcoming messages sent from the client and broadcast it to all other clients.

The below code creates a TCP socket and binds it to the server address, and then listens for upcoming connections:

import socket
from threading import Thread

# server's IP address
SERVER_PORT = 5002 # port we want to use
separator_token = "<SEP>" # we will use this to separate the client name & message

# initialize list/set of all connected client's sockets
client_sockets = set()
# create a TCP socket
s = socket.socket()
# make the port as reusable port
s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
# bind the socket to the address we specified
# listen for upcoming connections
print(f"[*] Listening as {SERVER_HOST}:{SERVER_PORT}")

Notice I've used "" as the server IP address. this means all IPv4 addresses on the local machine. You may wonder, why we don't just use localhost or "" ? Well, if the server has two IP addresses, let's say "" on a network and "" on another, then the server listens on both networks.

We're not yet accepting connections, as we didn't call accept() method, the below code finishes the server code recipe:

def listen_for_client(cs):
    This function keep listening for a message from `cs` socket
    Whenever a message is received, broadcast it to all other connected clients
    while True:
            # keep listening for a message from `cs` socket
            msg = cs.recv(1024).decode()
        except Exception as e:
            # client no longer connected
            # remove it from the set
            print(f"[!] Error: {e}")
            # if we received a message, replace the <SEP> 
            # token with ": " for nice printing
            msg = msg.replace(separator_token, ": ")
        # iterate over all connected sockets
        for client_socket in client_sockets:
            # and send the message

while True:
    # we keep listening for new connections all the time
    client_socket, client_address = s.accept()
    print(f"[+] {client_address} connected.")
    # add the new connected client to connected sockets
    # start a new thread that listens for each client's messages
    t = Thread(target=listen_for_client, args=(client_socket,))
    # make the thread daemon so it ends whenever the main thread ends
    t.daemon = True
    # start the thread

As mentioned previously, we add the connected client socket to our collection of our sockets, and then we start a new thread that executes our defined listen_for_client() function, which given a client socket, it waits for a message to be sent using recv() method, if so, then it sends it to all other connected clients.

Finally, let's close all sockets:

# close client sockets
for cs in client_sockets:
# close server socket

Alright, that's it for the server code, let's dive into the client code.

Client Code

The client does three basic operations:

  • Connects to the server.
  • Keep listening for messages coming from the server (must be a client sent a message to the server and the server broadcasted it) and print it to the console.
  • Waiting for user to input messages to send to the server.

Here's the code for the first operation:

import socket
import random
from threading import Thread
from datetime import datetime
from colorama import Fore, init, Back

# init colors

# set the available colors
colors = [Fore.BLUE, Fore.CYAN, Fore.GREEN, Fore.LIGHTBLACK_EX, 

# choose a random color for the client
client_color = random.choice(colors)

# server's IP address
# if the server is not on this machine, 
# put the private (network) IP address (e.g
SERVER_PORT = 5002 # server's port
separator_token = "<SEP>" # we will use this to separate the client name & message

# initialize TCP socket
s = socket.socket()
print(f"[*] Connecting to {SERVER_HOST}:{SERVER_PORT}...")
# connect to the server
print("[+] Connected.")

As a side operation, we also set a color for each client, you'll see it in the output. Also, let's set a name for each client, so we can distinguish between clients:

# prompt the client for a name
name = input("Enter your name: ")

The below code is responsible for the second operation; keep listening for messages from the server and print them to the console:

def listen_for_messages():
    while True:
        message = s.recv(1024).decode()
        print("\n" + message)

# make a thread that listens for messages to this client & print them
t = Thread(target=listen_for_messages)
# make the thread daemon so it ends whenever the main thread ends
t.daemon = True
# start the thread

We also want it to be in a separate thread, so we can do other things while listening for messages.

Now let's do the final task; waiting for user input for messages, and then send them to the server:

while True:
    # input message we want to send to the server
    to_send =  input()
    # a way to exit the program
    if to_send.lower() == 'q':
    # add the datetime, name & the color of the sender
    date_now ='%Y-%m-%d %H:%M:%S') 
    to_send = f"{client_color}[{date_now}] {name}{separator_token}{to_send}{Fore.RESET}"
    # finally, send the message

# close the socket

We add the client color, name and the current datetime to the message to be sent, we send the message using send() method, and we make a way to exit out of the problem, by just inputting 'q' character in the place of the message.


Alright, now that we finished both code recipes, let's make a demonstration. First, let's run one and only one server instance:

Server is listening for upcoming client connections

Awesome, the server is listening for upcoming client connections, let's try to run one client instance:

First client connectedNow the client is connected to the server and prompted for a username, to make sure it's connected, get back to the server console and you'll see indeed it's connected:

The Client is connected to the serverNote we're on localhost ( address for now, as it's the same machine, but if you want to connect from other machines in the same network, you can do that as well, just make sure to change SERVER_HOST in client code from to the server's private IP address.

Let's run another client so we can chat:

Second client is connected & chatting

Awesome, as you can see, each client have a color so we can distinguish between users, let's run a third client for fun:

Three clients chatting


Great, now every message sent from a particular client is sent to all other clients. Note the colors are changed whenever you re-execute the script.

Please check the full code so you can easily run them in your own!

I encourage you to add more features to this program. For example, you can make a notice to all users when a new client is connected!

For more Python sockets tutorials, check these:

Learn also: Logging in Python.

Happy Coding ♥

View Full Code
Sharing is caring!

Read Also

Comment panel