Python script that can downloads public and private profiles images and videos, like Gallery with photos or videos. It saves the data in the folder.

How it works:

  • Log in in instragram using selenium and navigate to the profile

  • Checking the availability of Instagram profile if it's private or existing

  • Creates a folder with the name of your choice

  • Gathering urls from images and videos

  • Using threads and multiprocessing improving the execution speed

My code:

from pathlib import Path
import requests
import time
from selenium import webdriver
from selenium.common.exceptions import NoSuchElementException, TimeoutException
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By
from multiprocessing.dummy import Pool
from concurrent.futures import ThreadPoolExecutor
from typing import *
import argparse
import shutil

class PrivateException(Exception):

class InstagramPV:
    MAX_WORKERS: int = 8
    N_PROCESSES: int = 8

    def __init__(self, username: str, password: str, folder: Path, profile_name: str):

        :param username: Username or E-mail for Log-in in Instagram
        :param password: Password for Log-in in Instagram
        :param folder: Folder name that will save the posts
        :param profile_name: The profile name that will search
        self.username = username
        self.password = password
        self.folder = folder
        self.http_base = requests.Session()
        self.profile_name = profile_name
        self.links: List[str] = []
        self.pictures: List[str] = []
        self.videos: List[str] = []
        self.url: str = 'https://www.instagram.com/{name}/'
        self.posts: int = 0
        self.driver = webdriver.Chrome()

    def __enter__(self):
        return self

    def __exit__(self, exc_type, exc_val, exc_tb):

    def check_availability(self) -> None:
        Checking Status code, Taking number of posts, Privacy and followed by viewer
        Raise Error if the Profile is private and not following by viewer
        :return: None
        search = self.http_base.get(self.url.format(name=self.profile_name), params={'__a': 1})

        load_and_check = search.json()
        self.posts = load_and_check.get('graphql').get('user').get('edge_owner_to_timeline_media').get('count')
        privacy = load_and_check.get('graphql').get('user').get('is_private')
        followed_by_viewer = load_and_check.get('graphql').get('user').get('followed_by_viewer')
        if privacy and not followed_by_viewer:
            raise PrivateException('[!] Account is private')

    def create_folder(self) -> None:
        """Create the folder name"""

    def login(self) -> None:
        """Login To Instagram"""
        WebDriverWait(self.driver, 10).until(EC.presence_of_element_located((By.TAG_NAME, 'form')))
        submit = self.driver.find_element_by_tag_name('form')

        """Check For Invalid Credentials"""
            var_error = WebDriverWait(self.driver, 4).until(EC.presence_of_element_located((By.CLASS_NAME, 'eiCW-')))
            raise ValueError(var_error.text)
        except TimeoutException:

            """Close Notifications"""
            notifications = WebDriverWait(self.driver, 20).until(
                EC.presence_of_element_located((By.XPATH, '//button[text()="Not Now"]')))
        except NoSuchElementException:

        """Taking cookies"""
        cookies = {
            cookie['name']: cookie['value']
            for cookie in self.driver.get_cookies()


        """Check for availability"""



    def posts_urls(self) -> None:
        """Taking the URLs from posts and appending in self.links"""
        elements = self.driver.find_elements_by_xpath('//a[@href]')
        for elem in elements:
            urls = elem.get_attribute('href')
            if urls not in self.links and 'p' in urls.split('/'):

    def scroll_down(self) -> None:
        """Scrolling down the page and taking the URLs"""
        last_height = self.driver.execute_script('return document.body.scrollHeight')
        while True:
            self.driver.execute_script('window.scrollTo(0, document.body.scrollHeight);')
            new_height = self.driver.execute_script("return document.body.scrollHeight")
            if new_height == last_height:
            last_height = new_height

    def submit_links(self) -> None:
        """Gathering Images and Videos and pass to function <fetch_url> Using ThreadPoolExecutor"""


        print('[!] Ready for video - images'.title())
        print(f'[*] extracting {len(self.links)} posts , please wait...'.title())

        with ThreadPoolExecutor(max_workers=self.MAX_WORKERS) as executor:
            for link in self.links:
                executor.submit(self.fetch_url, link)

    def get_fields(self, nodes: Dict, *keys) -> Any:
        :param nodes: The json data from the link using only the first two keys 'graphql' and 'shortcode_media'
        :param keys: Keys that will be add to the nodes and will have the results of 'type' or 'URL'
        :return: The value of the key <fields>
        fields = nodes['graphql']['shortcode_media']
        for key in keys:
            fields = fields[key]
        return fields

    def fetch_url(self, url: str) -> None:
        This function extracts images and videos
        :param url: Taking the url
        :return None

        logging_page_id = self.http_base.get(url, params={'__a': 1}).json()
        if self.get_fields(logging_page_id, '__typename') == 'GraphImage':
            image_url = self.get_fields(logging_page_id, 'display_url')

        elif self.get_fields(logging_page_id, '__typename') == 'GraphVideo':
            video_url = self.get_fields(logging_page_id, 'video_url')

        elif self.get_fields(logging_page_id, '__typename') == 'GraphSidecar':
            for sidecar in self.get_fields(logging_page_id, 'edge_sidecar_to_children', 'edges'):
                if sidecar['node']['__typename'] == 'GraphImage':
                    image_url = sidecar['node']['display_url']
                    video_url = sidecar['node']['video_url']
            print(f'Warning {url}: has unknown type of {self.get_fields(logging_page_id,"__typename")}')

    def download_video(self, new_videos: Tuple[int, str]) -> None:
        Saving the video content
        :param new_videos: Tuple[int,str]
        :return: None
        number, link = new_videos

        with open(self.folder / f'Video{number}.mp4', 'wb') as f:
            content_of_video = self.http_base.get(link, stream=True).raw
            shutil.copyfileobj(content_of_video, f)

    def images_download(self, new_pictures: Tuple[int, str]) -> None:
        Saving the picture content
        :param new_pictures: Tuple[int, str]
        :return: None

        number, link = new_pictures
        with open(self.folder / f'Image{number}.jpg', 'wb') as f:
            content_of_picture = self.http_base.get(link, stream=True).raw
            shutil.copyfileobj(content_of_picture, f)

    def downloading_video_images(self) -> None:
        """Using multiprocessing for Saving Images and Videos"""
        print('[*] ready for saving images and videos!'.title())
        picture_data = enumerate(self.pictures)
        video_data = enumerate(self.videos)
        pool = Pool(self.N_PROCESSES)
        pool.map(self.images_download, picture_data)
        pool.map(self.download_video, video_data)
        print('[+] Done')

def main():
    parser = argparse.ArgumentParser()

    parser.add_argument('-U', '--username', help='Username or your email of your account', action='store',
    parser.add_argument('-P', '--password', help='Password of your account', action='store', required=True)
    parser.add_argument('-F', '--filename', help='Filename for storing data', action='store', required=True)
    parser.add_argument('-T', '--target', help='Profile name to search', action='store', required=True)
    args = parser.parse_args()

    with InstagramPV(args.username, args.password, Path(args.filename), args.target) as pv:

if __name__ == '__main__':

Usage: myfile.py -U [email protected] -P mypassword -F Mynamefile -T stackoverjoke

My previous comparative review tag:Download pictures (or videos) from Instagram using Selenium


1 Answer 1


More constants


    self.url: str = 'https://www.instagram.com/{name}/'

appears to be a constant, so it can join the others in class scope. While you're doing that, you can also pull the URL from self.driver.get('https://www.instagram.com/accounts/login') into a constant; and also pull the base URL out. In other words:

class InstagramPV:
    MAX_WORKERS: int = 8
    N_PROCESSES: int = 8

    BASE_URL = 'https://www.instagram.com/'
    PROFILE_URL_FMT = BASE_URL + '{name}/'
    LOGIN_URL = BASE_URL + 'accounts/login'

Nested get



won't actually do what you want, which is a fail-safe object traversal. For that you need to provide defaults that are empty dictionaries:

self.posts = (
    load_and_check.get('graphql', {})
    .get('user', {})
    .get('edge_owner_to_timeline_media', {})

Also, the first part should be factored out into its own temporary variable, since it's used three times:

user = (
    load_and_check.get('graphql', {})
    .get('user', {})

Methods for reuse

self.driver.execute_script("return document.body.scrollHeight")

should be factored out into a new method for re-use.

Static function


def get_fields(self, nodes: Dict, *keys) -> Any:
    :param nodes: The json data from the link using only the first two keys 'graphql' and 'shortcode_media'
    :param keys: Keys that will be add to the nodes and will have the results of 'type' or 'URL'
    :return: The value of the key <fields>
    fields = nodes['graphql']['shortcode_media']
    for key in keys:
        fields = fields[key]
    return fields

doesn't ever use self, which is a big clue that it doesn't belong as an instance method. You should just make it a @staticmethod. The only reason I don't recommend it moving to global scope is that it still has knowledge of the Instagram data format, with its reference to graphql.

Dictionary traversal

The loop in get_fields can be replaced with a call to functools.reduce(dict.get, keys, media). Also, keys - even though it is a variadic argument - can still receive a type hint, and should be Iterable[str]. nodes itself, if you don't know a lot about the structure of the dictionary, can still be narrowed to nodes: Dict[str, Any].

Context manager for response

Now that you're using the streaming interface for Requests (nice!), it's more important that you use the response object as a context manager. For more information read https://github.com/psf/requests/issues/4136


with open(self.folder / f'Video{number}.mp4', 'wb') as f, \
     self.http_base.get(link, stream=True) as response:
    shutil.copyfileobj(response.raw, f)
  • \$\begingroup\$ Questions : 1) Methods for reuse should i put this in a variable for reuse? 2) Static function the field returns 'str' or 'list' , can Any be Any[str,list] ? 3) Staticmethods should be in the end of class or doesn't matter ? \$\endgroup\$
    – AlexDotis
    Commented Apr 12, 2020 at 18:12
  • \$\begingroup\$ 1. If it never changes, then yes, a variable is better than a method. \$\endgroup\$
    – Reinderien
    Commented Apr 12, 2020 at 18:13
  • \$\begingroup\$ 2. This is a job for Union! If you know it may be str or list, then Union[str, list]. \$\endgroup\$
    – Reinderien
    Commented Apr 12, 2020 at 18:14
  • \$\begingroup\$ Let us continue this discussion in chat. \$\endgroup\$
    – Reinderien
    Commented Apr 12, 2020 at 18:55
  • 1
    \$\begingroup\$ To summarize the reason that I will stand by this suggestion: being Pythonic is about simplicity, and this is simple; functools should be avoided when there is a simpler tool that does the same job, but there is not. stackoverflow.com/a/36584550/313768 remains (in my opinion, rightfully) an acceptable answer. In many, many other cases, reduce is a bad idea. \$\endgroup\$
    – Reinderien
    Commented Apr 12, 2020 at 19:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.