Python Telegram bot api.

License: GPL-2.0

Language: Python

Keywords: bot-api, python-api, telegram, telegram-bot, telegram-bot-api


A simple, but extensible Python implementation for the Telegram Bot API.

Download Month Build Status

Getting started.

This API is tested with Python 2.6, Python 2.7, Python 3.4, Pypy and Pypy 3. There are two ways to install the library:

  • Installation using pip (a Python package manager)*:
$ pip install pyTelegramBotAPI
  • Installation from source (requires git):
$ git clone
$ cd pyTelegramBotAPI
$ python install

It is generally recommended to use the first option.

*While the API is production-ready, it is still under development and it has regular updates, do not forget to update it regularly by calling pip install pytelegrambotapi --upgrade

Writing your first bot


It is presumed that you have obtained an API token with @BotFather. We will call this token TOKEN. Furthermore, you have basic knowledge of the Python programming language and more importantly the Telegram Bot API.

A simple echo bot

The TeleBot class (defined in encapsulates all API calls in a single class. It provides functions such as send_xyz (send_message, send_document etc.) and several ways to listen for incoming messages.

Create a file called Then, open the file and create an instance of the TeleBot class.

import telebot

bot = telebot.TeleBot("TOKEN")

Note: Make sure to actually replace TOKEN with your own API token.

After that declaration, we need to register some so-called message handlers. Message handlers define filters which a message must pass. If a message passes the filter, the decorated function is called and the incoming message is passed as an argument.

Let's define a message handler which handles incoming /start and /help commands.

@bot.message_handler(commands=['start', 'help'])
def send_welcome(message):
	bot.reply_to(message, "Howdy, how are you doing?")

A function which is decorated by a message handler can have an arbitrary name, however, it must have only one parameter (the message).

Let's add another handler:

@bot.message_handler(func=lambda m: True)
def echo_all(message):
	bot.reply_to(message, message.text)

This one echoes all incoming text messages back to the sender. It uses a lambda function to test a message. If the lambda returns True, the message is handled by the decorated function. Since we want all messages to be handled by this function, we simply always return True.

Note: all handlers are tested in the order in which they were declared

We now have a basic bot which replies a static message to "/start" and "/help" commands and which echoes the rest of the sent messages. To start the bot, add the following to our source file:


Alright, that's it! Our source file now looks like this:

import telebot

bot = telebot.TeleBot("TOKEN")

@bot.message_handler(commands=['start', 'help'])
def send_welcome(message):
	bot.reply_to(message, "Howdy, how are you doing?")

@bot.message_handler(func=lambda message: True)
def echo_all(message):
	bot.reply_to(message, message.text)


To start the bot, simply open up a terminal and enter python to run the bot! Test it by sending commands ('/start' and '/help') and arbitrary text messages.

General API Documentation


All types are defined in They are all completely in line with the Telegram API's definition of the types, except for the Message's from field, which is renamed to from_user (because from is a Python reserved token). Thus, attributes such as message_id can be accessed directly with message.message_id. Note that can be either an instance of User or GroupChat (see How can I distinguish a User and a GroupChat in

The Message object also has a content_typeattribute, which defines the type of the Message. content_type can be one of the following strings: text, audio, document, photo, sticker, video, video_note, voice, location, contact, new_chat_members, left_chat_member, new_chat_title, new_chat_photo, delete_chat_photo, group_chat_created, supergroup_chat_created, channel_chat_created, migrate_to_chat_id, migrate_from_chat_id, pinned_message.

You can use some types in one function. Example:

content_types=["text", "sticker", "pinned_message", "photo", "audio"]


All API methods are located in the TeleBot class. They are renamed to follow common Python naming conventions. E.g. getMe is renamed to get_me and sendMessage to send_message.

General use of the API

Outlined below are some general use cases of the API.

Message handlers

A message handler is a function that is decorated with the message_handler decorator of a TeleBot instance. Message handlers consist of one or multiple filters. Each filter much return True for a certain message in order for a message handler to become eligible to handle that message. A message handler is declared in the following way (provided bot is an instance of TeleBot):

def function_name(message):
	bot.reply_to(message, "This is a message handler")

function_name is not bound to any restrictions. Any function name is permitted with message handlers. The function must accept at most one argument, which will be the message that the function must handle. filters is a list of keyword arguments. A filter is declared in the following manner: name=argument. One handler may have multiple filters. TeleBot supports the following filters:

name argument(s) Condition
content_types list of strings (default ['text']) True if message.content_type is in the list of strings.
regexp a regular expression as a string True if returns True and message.content_type == 'text' (See Python Regular Expressions
commands list of strings True if message.content_type == 'text' and message.text starts with a command that is in the list of strings.
func a function (lambda or function reference) True if the lambda or function reference returns True

Here are some examples of using the filters and message handlers:

import telebot
bot = telebot.TeleBot("TOKEN")

# Handles all text messages that contains the commands '/start' or '/help'.
@bot.message_handler(commands=['start', 'help'])
def handle_start_help(message):

# Handles all sent documents and audio files
@bot.message_handler(content_types=['document', 'audio'])
def handle_docs_audio(message):

# Handles all text messages that match the regular expression
def handle_message(message):

#Handles all messages for which the lambda returns True
@bot.message_handler(func=lambda message: message.document.mime_type == 'text/plain', content_types=['document'])
def handle_text_doc(message):

#Which could also be defined as:
def test_message(message):
	return message.document.mime_type == 'text/plain'

@bot.message_handler(func=test_message, content_types=['document'])
def handle_text_doc(message)

# Handlers can be stacked to create a function which will be called if either message_handler is eligible
# This handler will be called if the message starts with '/hello' OR is some emoji
@bot.message_handler(func=lambda msg: msg.text.encode("utf-8") == SOME_FANCY_EMOJI)
def send_something(message):

Important: all handlers are tested in the order in which they were declared

Edited Message handlers

Same as Message handlers


Same as Message handlers


Same as Message handlers

Callback Query Handler

In bot2.0 update. You can get callback_query in update object. In telebot use callback_query_handler to process callback_querys.

@bot.callback_query_handler(func=lambda call: True)
def  test_callback(call):


import telebot

TOKEN = '<token_string>'
tb = telebot.TeleBot(TOKEN)	#create a new Telegram Bot object

# Upon calling this function, TeleBot starts polling the Telegram servers for new messages.
# - none_stop: True/False (default False) - Don't stop polling when receiving an error from the Telegram servers
# - interval: True/False (default False) - The interval between polling requests
#           Note: Editing this parameter harms the bot's response time
# - timeout: integer (default 20) - Timeout in seconds for long polling.
tb.polling(none_stop=False, interval=0, timeout=20)

# getMe
user = tb.get_me()

# setWebhook
tb.set_webhook(url="", certificate=open('mycert.pem'))
# unset webhook

# getUpdates
updates = tb.get_updates()
updates = tb.get_updates(1234,100,20) #get_Updates(offset, limit, timeout):

# sendMessage
tb.send_message(chatid, text)

# forwardMessage
tb.forward_message(to_chat_id, from_chat_id, message_id)

# All send_xyz functions which can take a file as an argument, can also take a file_id instead of a file.
# sendPhoto
photo = open('/tmp/photo.png', 'rb')
tb.send_photo(chat_id, photo)
tb.send_photo(chat_id, "FILEID")

# sendAudio
audio = open('/tmp/audio.mp3', 'rb')
tb.send_audio(chat_id, audio)
tb.send_audio(chat_id, "FILEID")

## sendAudio with duration, performer and title.
tb.send_audio(CHAT_ID, file_data, 1, 'eternnoir', 'pyTelegram')

# sendVoice
voice = open('/tmp/voice.ogg', 'rb')
tb.send_voice(chat_id, voice)
tb.send_voice(chat_id, "FILEID")

# sendDocument
doc = open('/tmp/file.txt', 'rb')
tb.send_document(chat_id, doc)
tb.send_document(chat_id, "FILEID")

# sendSticker
sti = open('/tmp/sti.webp', 'rb')
tb.send_sticker(chat_id, sti)
tb.send_sticker(chat_id, "FILEID")

# sendVideo
video = open('/tmp/video.mp4', 'rb')
tb.send_video(chat_id, video)
tb.send_video(chat_id, "FILEID")

# sendVideoNote
videonote = open('/tmp/videonote.mp4', 'rb')
tb.send_video_note(chat_id, videonote)
tb.send_video_note(chat_id, "FILEID")

# sendLocation
tb.send_location(chat_id, lat, lon)

# sendChatAction
# action_string can be one of the following strings: 'typing', 'upload_photo', 'record_video', 'upload_video',
# 'record_audio', 'upload_audio', 'upload_document' or 'find_location'.
tb.send_chat_action(chat_id, action_string)

# getFile
# Downloading a file is straightforward
# Returns a File object
import requests
file_info = tb.get_file(file_id)

file = requests.get('{0}/{1}'.format(API_TOKEN, file_info.file_path))

Reply markup

All send_xyz functions of TeleBot take an optional reply_markup argument. This argument must be an instance of ReplyKeyboardMarkup, ReplyKeyboardRemove or ForceReply, which are defined in

from telebot import types

# Using the ReplyKeyboardMarkup class
# It's constructor can take the following optional arguments:
# - resize_keyboard: True/False (default False)
# - one_time_keyboard: True/False (default False)
# - selective: True/False (default False)
# - row_width: integer (default 3)
# row_width is used in combination with the add() function.
# It defines how many buttons are fit on each row before continuing on the next row.
markup = types.ReplyKeyboardMarkup(row_width=2)
itembtn1 = types.KeyboardButton('a')
itembtn2 = types.KeyboardButton('v')
itembtn3 = types.KeyboardButton('d')
markup.add(itembtn1, itembtn2, itembtn3)
tb.send_message(chat_id, "Choose one letter:", reply_markup=markup)

# or add KeyboardButton one row at a time:
markup = types.ReplyKeyboardMarkup()
itembtna = types.KeyboardButton('a')
itembtnv = types.KeyboardButton('v')
itembtnc = types.KeyboardButton('c')
itembtnd = types.KeyboardButton('d')
itembtne = types.KeyboardButton('e')
markup.row(itembtna, itembtnv)
markup.row(itembtnc, itembtnd, itembtne)
tb.send_message(chat_id, "Choose one letter:", reply_markup=markup)

The last example yields this result:


# ReplyKeyboardRemove: hides a previously sent ReplyKeyboardMarkup
# Takes an optional selective argument (True/False, default False)
markup = types.ReplyKeyboardRemove(selective=False)
tb.send_message(chat_id, message, reply_markup=markup)
# ForceReply: forces a user to reply to a message
# Takes an optional selective argument (True/False, default False)
markup = types.ForceReply(selective=False)
tb.send_message(chat_id, "Send me another word:", reply_markup=markup)



Inline Mode

More information about Inline mode.


Now, you can use inline_handler to get inline_query in telebot.

@bot.inline_handler(lambda query: query.query == 'text')
def query_text(inline_query):
    # Query message is text


Use chosen_inline_handler to get chosen_inline_result in telebot. Don't forgot add the /setinlinefeedback command for @Botfather.

More information : collecting-feedback

@bot.chosen_inline_handler(func=lambda chosen_inline_result: True)
def test_chosen(chosen_inline_result):
    # Process all chosen_inline_result.


@bot.inline_handler(lambda query: query.query == 'text')
def query_text(inline_query):
        r = types.InlineQueryResultArticle('1', 'Result', types.InputTextMessageContent('Result message.'))
        r2 = types.InlineQueryResultArticle('2', 'Result2', types.InputTextMessageContent('Result message2.'))
        bot.answer_inline_query(, [r, r2])
    except Exception as e:

Working with entities:

This object represents one special entity in a text message. For example, hashtags, usernames, URLs, etc. Attributes:

  • type
  • url
  • offset
  • length
  • user

Here's an Example:message.entities[num].<attribute>
Here num is the entity number or order of entity in a reply, for if incase there are multiple entities in the reply/message.
message.entities returns a list of entities object.
message.entities[0].type would give the type of the first entity
Refer Bot Api for extra details

Advanced use of the API

Asynchronous delivery of messages

There exists an implementation of TeleBot which executes all send_xyz and the get_me functions asynchronously. This can speed up you bot significantly, but it has unwanted side effects if used without caution. To enable this behaviour, create an instance of AsyncTeleBot instead of TeleBot.

tb = telebot.AsyncTeleBot("TOKEN")

Now, every function that calls the Telegram API is executed in a separate Thread. The functions are modified to return an AsyncTask instance (defined in Using AsyncTeleBot allows you to do the following:

import telebot

tb = telebot.AsyncTeleBot("TOKEN")
task = tb.get_me() # Execute an API call
# Do some other operations...
a = 0
for a in range(100):
	a += 10

result = task.wait() # Get the result of the execution

Note: if you execute send_xyz functions after eachother without calling wait(), the order in which messages are delivered might be wrong.

Sending large text messages

Sometimes you must send messages that exceed 5000 characters. The Telegram API can not handle that many characters in one request, so we need to split the message in multiples. Here is how to do that using the API:

from telebot import util
large_text = open("large_text.txt", "rb").read()

# Split the text each 3000 characters.
# split_string returns a list with the splitted text.
splitted_text = util.split_string(large_text, 3000)
for text in splitted_text:
	tb.send_message(chat_id, text)

Controlling the amount of Threads used by TeleBot

The TeleBot constructor takes the following optional arguments:

  • threaded: True/False (default True). A flag to indicate whether TeleBot should execute message handlers on it's polling Thread.

The listener mechanism

As an alternative to the message handlers, one can also register a function as a listener to TeleBot. Example:

def handle_messages(messages):
	for message in messages:
		# Do something with the message
		bot.reply_to(message, 'Hi')


Using web hooks

When using webhooks telegram sends one Update per call, for processing it you should call process_new_messages([update.message]) when you recieve it.

There are some examples using webhooks in the examples/webhook_examples directory.


You can use the Telebot module logger to log debug info about Telebot. Use telebot.logger to get the logger of the TeleBot module. It is possible to add custom logging Handlers to the logger. Refer to the Python logging module page for more info.

import logging

logger = telebot.logger
telebot.logger.setLevel(logging.DEBUG) # Outputs debug messages to console.


You can use proxy for request. apihelper.proxy object will use by call requests proxies argument.

from telebot import apihelper

apihelper.proxy = {'http':''}

If you want to use socket5 proxy you need install dependency pip install requests[socks] and make sure, that you have the latest version of gunicorn, PySocks, pyTelegramBotAPI, requests and urllib3.

apihelper.proxy = {'https':'socks5://userproxy:password@proxy_address:port'}


Bot 2.0

April 9,2016 Telegram release new bot 2.0 API, which has a drastic revision especially for the change of method's interface.If you want to update to the latest version, please make sure you've switched bot's code to bot 2.0 method interface.

More information about pyTelegramBotAPI support bot2.0

How can I distinguish a User and a GroupChat in

Telegram Bot API support new type Chat for

  • Check the type attribute in Chat object:
if == “private”:
	# private chat message

if == “group”:
	# group chat message

if == “supergroup”:
	# supergroup chat message

if == “channel”:
	# channel message

The Telegram Chat Group

Get help. Discuss. Chat.

More examples

Bots using this API

Want to have your bot listed here? Send a Telegram message to @eternnoir or @pevdh.

Project Statistics

Sourcerank 11
Repository Size 1.44 MB
Stars 1,983
Forks 561
Watchers 140
Open issues 161
Dependencies 7
Contributors 71
Tags 26
Last updated
Last pushed

Top Contributors See all

FrankWang Pieter Evgeny Petrov WaffleWafflerov Sergey sviat9440 Sergey Max Dmitry user Leandro Späth Andrew Developer JekaFST DrBomb Ihor Polyakov the31k Badiboy Dmytro Striletskyi Sergio Conde Gómez Gabriel R F

Packages Referencing this Repo

Python Telegram bot api.
Latest release 3.6.6 - Updated - 1.98K stars

Recent Tags See all

3.6.6 September 07, 2018
3.6.5 August 03, 2018
3.6.4 July 31, 2018
3.6.3 May 14, 2018
3.6.2 March 23, 2018
3.6.1 March 10, 2018
3.6.0 March 02, 2018
3.5.2 January 24, 2018
3.5.1 November 30, 2017
latest_release November 29, 2017
3.5.0 November 14, 2017
3.2.0 August 06, 2017
3.1.0 July 01, 2017
2.2.3 October 31, 2016
2.2.0 October 29, 2016

Interesting Forks See all

Python Telegram bot api.
Python - Last pushed - 3 stars - 4 forks
Python Telegram bot api.
Python - GPL-2.0 - Updated - 2 stars
Python Telegram bot api.
Python - Last pushed - 1 stars
Python Telegram bot api.
Python - Updated - 1 stars
Python Telegram bot api.
Python - Last pushed - 1 stars

Something wrong with this page? Make a suggestion

Last synced: 2018-08-03 00:54:58 UTC

Login to resync this repository