Implemented in Python using the Selenium module.
Think this tool is worth supporting? Head over to https://github.com/timgrossmann/InstaPy/wiki/How-to-Contribute to find out how you can help. Become a part of InstaPy!
Have an issue Head over to https://github.com/timgrossmann/InstaPy/wiki/Reporting-An-Issue to find out how to report this to us and get help.
Disclaimer: Please Note that this is a research project. I am by no means responsible for any usage of this tool. Use on your own behalf. I’m also not responsible if your accounts get banned due to extensive use of this tool.
Newsletter: SignUp for the Newsletter here!
Listen to the "Talk Python to me"-Episode | Support InstaPy!
- Getting Started
- InstaPy Available Features
- Commenting
- Following
- Following by a list
- Follow someone else's followers
- Follow users that someone else is following
- Follow someone else's followers/following
- Interact with specific users
- Interact with users that someone else is following
- Interact with someone else's followers
- Unfollowing
- Don't unfollow active users
- Interactions based on the number of followers a user has
- Like by Locations
- Like by Tags
- Like by Feeds
- Restricting Likes
- Ignoring Users
- Ignoring Restrictions
- Excluding friends
- Blacklist Campaign
- Smart Hashtags
- Follow/Unfollow/exclude not working?
- Third Party InstaPy GUI for Windows
- Switching to Firefox
- Emoji Support
- Clarifai ImageAPI
- Running on a Server
- Running on a Headless Browser
- Running with Docker microservices manual
- Running all-in-one with Docker (obsolete)
- Automate with cron
- Automate with Schedule
- Extra Informations
Setting up InstaPy at Digital Ocean (for Debian)
1. git clone https://github.com/timgrossmann/InstaPy.git
2. cd InstaPy
3. pip install .
or
3. python setup.py install- Download
chromedriverfor your system from here. And put it in/assetsfolder.
Basic setup is a good way to test the tool. At project root folder open quickstart.py and update with your username and password.
frominstapyimportInstaPyinsta_username=''insta_password=''# if you want to run this script on a server, # simply add nogui=True to the InstaPy() constructorsession=InstaPy(username=insta_username, password=insta_password)
session.login()
# set up all the settingssession.set_upper_follower_count(limit=2500)
session.set_do_comment(True, percentage=10)
session.set_comments(['aMEIzing!', 'So much fun!!', 'Nicey!'])
session.set_dont_include(['friend1', 'friend2', 'friend3'])
session.set_dont_like(['pizza', 'girl'])
# do the actual likingsession.like_by_tags(['natgeo', 'world'], amount=100)
# end the bot sessionsession.end()Execute it:
$ python quickstart.py1. Official Cross Platform GUI
2. Third Party InstaPy GUI for Windows
3. Session scheduling with Telegram
# default enabled=False, ~ every 4th image will be commented onsession.set_do_comment(enabled=True, percentage=25)
session.set_comments(['Awesome', 'Really Cool', 'I like your stuff'])
# you can also set comments for specific media types (Photo / Video)session.set_comments(['Nice shot!'], media='Photo')
session.set_comments(['Great Video!'], media='Video')# default enabled=False, follows ~ 10% of the users from the images, times=1 # (only follows a user once (if unfollowed again))session.set_do_follow(enabled=True, percentage=10, times=2)# follows each account from a list of instagram nicknames (only follows a user# once (if unfollowed again)) would be useful for the precise targeting. # For example, if one needs to get followbacks from followers of a chosen # account/group of accounts.accs= ['therock','natgeo']
session.follow_by_list(accs, times=1)# Follows the followers of each given user# The usernames can be either a list or a string# The amount is for each account, in this case 30 users will be followed# If randomize is false it will pick in a top-down fashionsession.follow_user_followers(['friend1', 'friend2', 'friend3'], amount=10, randomize=False)
# default sleep_delay=600 (10min) for every 10 user following, in this case# sleep for 60 seconds session.follow_user_followers(['friend1', 'friend2', 'friend3'], amount=10, randomize=False, sleep_delay=60)# Follows the people that a given users are following# The usernames can be either a list or a string# The amount is for each account, in this case 30 users will be followed# If randomize is false it will pick in a top-down fashionsession.follow_user_following(['friend1', 'friend2', 'friend3'], amount=10, randomize=False)
# default sleep_delay=600 (10min) for every 10 user following, in this case# sleep for 60 secondssession.follow_user_following(['friend1', 'friend2', 'friend3'], amount=10, randomize=False, sleep_delay=60)# For 50% of the 30 newly followed, move to their profile# and randomly choose 5 pictures to be liked.# Take into account the other set options like the comment rate# and the filtering for inappropriate words or userssession.set_user_interact(amount=5, randomize=True, percentage=50, media='Photo')
session.follow_user_followers(['friend1', 'friend2', 'friend3'], amount=10, randomize=False, interact=True)# Interact with specific users# set_do_like, set_do_comment, set_do_follow are applicablesession.set_do_follow(enabled=False, percentage=50)
session.set_comments(["Cool", "Super!"])
session.set_do_comment(enabled=True, percentage=80)
session.set_do_like(True, percentage=70)
session.interact_by_users(['user1', 'user2', 'user3'], amount=5, randomize=True, media='Photo')# Interact with the people that a given user is following# set_do_comment, set_do_follow and set_do_like are applicablesession.set_user_interact(amount=5, randomize=True, percentage=50, media='Photo')
session.set_do_follow(enabled=False, percentage=70)
session.set_do_like(enabled=False, percentage=70)
session.set_comments(["Cool", "Super!"])
session.set_do_comment(enabled=True, percentage=80)
session.interact_user_following(['natgeo'], amount=10, randomize=True)# Interact with the people that a given user is following# set_do_comment, set_do_follow and set_do_like are applicablesession.set_user_interact(amount=5, randomize=True, percentage=50, media='Photo')
session.set_do_follow(enabled=False, percentage=70)
session.set_do_like(enabled=False, percentage=70)
session.set_comments(["Cool", "Super!"])
session.set_do_comment(enabled=True, percentage=80)
session.interact_user_followers(['natgeo'], amount=10, randomize=True)# unfollows 10 of the accounts you're following -> instagram will only # unfollow 10 before you'll be 'blocked for 10 minutes' (if you enter a # higher number than 10 it will unfollow 10, then wait 10 minutes and will # continue then).# You can choose to only unfollow the user that Insta has followed by adding# onlyInstapyFollowed = True otherwise it will unfollow all users# You can choose unfollow method as FIFO (First-Input-First-Output) or # LIFO (Last-Input-First-Output). The default is FIFO method. # onlyInstapyMethod is using only when onlyInstapyFollowed = True# sleep_delay sets the time it will sleep every 10 profile unfollow, default# is 10minsession.unfollow_users(amount=10, onlyInstapyFollowed=True, onlyInstapyMethod='FIFO', sleep_delay=60 )
# You can only unfollow user that won't follow you back by adding# onlyNotFollowMe = True it still only support on profile following # you should disable onlyInstapyFollowed when use this session.unfollow_users(amount=10, onlyNotFollowMe=True, sleep_delay=60)# Prevents unfollow followers who have liked one of your latest 5 postssession.set_dont_unfollow_active_users(enabled=True, posts=5)# This is used to check the number of followers a user has and if this number # exceeds the number set then no further interaction happenssession.set_upper_follower_count(limit=250)# This is used to check the number of followers a user has and if this number# does not pass the number set then no further interaction happenssession.set_lower_follower_count(limit=1)session.like_by_locations(['224442573/salton-sea/'], amount=100)
# orsession.like_by_locations(['224442573'], amount=100)
# or include media entities from top posts sectionsession.like_by_locations(['224442573'], amount=5, skip_top_posts=False)You can find locations for the like_by_locations function by:
- Browsing https://www.instagram.com/explore/locations/
- Regular instagram search.
Example:
- Search 'Salton Sea' and select the result with a location icon
- The url is: https://www.instagram.com/explore/locations/224442573/salton-sea/
- Use everything after 'locations/' or just the number
# Like posts based on hashtagssession.like_by_tags(['natgeo', 'world'], amount=10)# This is used to perform likes on your own feeds# amount=100 specifies how many total likes you want to perform# randomize=True randomly skips posts to be liked on your feed# unfollow=True unfollows the author of a post which was considered # inappropriate interact=True visits the author's profile page of a # certain post and likes a given number of his pictures, then returns to feedsession.like_by_feed(amount=100, randomize=True, unfollow=True, interact=True)# Controls your interactions by campaigns.# ex. this week InstaPy will like and comment interacting by campaign called# 'soccer', next time InstaPy runs, it will not interact again with users in # blacklist# In general, this means that once we turn off the soccer_campaign again, InstaPy# will have no track of the people it interacted with about soccer.# This will help you target people only once but several times for different campaignssession.set_blacklist(enabled=True, campaign='soccer_campaign')
session.set_do_comment(True, percentage=50)
session.set_comments(['Neymar is better than CR7', 'Soccer is cool'])
session.like_by_tags(['soccer', 'cr7', 'neymar'], amount=100, media='Photo')# Generate smart hashtags based on https://displaypurposes.com ranking,# banned and spammy tags are filtered out.# (limit) defines amount limit of generated hashtags by hashtag# (sort) sort generated hashtag list 'top' and 'random' are available# (log_tags) shows generated hashtags before use it# (use_smart_hashtags) activates like_by_tag to use smart hashtagssession.set_smart_hashtags(['cycling', 'roadbike'], limit=3, sort='top', log_tags=True)
session.like_by_tags(amount=10, use_smart_hashtags=True)session.set_dont_like(['#exactmatch', '[startswith', ']endswith', 'broadmatch']).set_dont_like searches the description and owner comments for hashtags and
won't like the image if one of those hashtags are in there
You have 4 options to exclude posts from your InstaPy session:
- words starting with
#will match only exact hashtags (e. g.#catmatches#cat, but not#catpic) - words starting with
[will match all hashtags starting with your word (e. g.[catmatches#catpic,#caturdayand so on) - words starting with
]will match all hashtags ending with your word (e. g.]catmatches#mycat,#instacatand so on) - words without these prefixes will match all hashtags that contain your word regardless if it is placed at the beginning, middle or end of the hashtag (e. g.
catwill match#cat,#mycat,#caturday,#rainingcatsanddogsand so on)
# completely ignore liking images from certain userssession.set_ignore_users(['random_user', 'another_username'])# will ignore the don't like if the description contains# one of the given wordssession.set_ignore_if_contains(['glutenfree', 'french', 'tasty'])# will prevent commenting on and unfollowing your good friends (the images will # still be liked)session.set_dont_include(['friend1', 'friend2', 'friend3'])If you notice that one or more of the above functionalities are not working as expected - e.g. you have specified:
session.set_do_follow(enabled=True, percentage=10, times=2)but none of the profiles are being followed - or any such functionality is misbehaving - then one thing you should check is the position/order of such methods in your script. Essentially, all the set_* methods have to be before like_by_tags or like_by_locations or unfollow. This is also implicit in all the exmples and quickstart.py
Chrome is the default browser, but InstaPy provides support for Firefox as well.
session=InstaPy(username=insta_username, password=insta_password, use_firefox=True)To use an emoji just add an u in front of the opening apostrophe:
session.set_comments([u'This post is 🔥',u'More emojis are always better 💯',u'I love your posts 😍😍😍']);
# or
session.set_comments([u'Emoji text codes are also supported :100: :thumbsup: :thumbs_up: \u2764 💯💯']);
Emoji text codes are implemented using 2 different naming codes. A complete list of emojis codes can be found on the Python Emoji Github, but you can use the alternate shorted naming scheme found for Emoji text codes here. Note: Every Emoji has not been tested. Please report any inconsistencies.
Legacy Emoji Support
You can still use Unicode strings in your comments, but there are some limitations.
You can use only Unicode characters with no more than 4 characters and you have to use the unicode code (e. g.
\u1234). You find a list of emoji with unicode codes on Wikipedia, but there is also a list of working emoji in/assetsYou have to convert your comment to Unicode. This can safely be done by adding an u in front of the opening apostrophe:
u'\u1234 some comment'
Note: Head over to https://developer.clarifai.com/signup/ and create a free account, once you’re logged in go to https://developer.clarifai.com/account/applications/ and create a new application. You can find the client ID and Secret there. You get 5000 API-calls free/month.
If you want the script to get your CLARIFAI_API_KEY for your environment, you can do:
export CLARIFAI_API_KEY="<API KEY>"
session.set_do_comment(True, percentage=10)
session.set_comments(['Cool!', 'Awesome!', 'Nice!'])
session.set_use_clarifai(enabled=True)
session.clarifai_check_img_for(['nsfw'])
session.clarifai_check_img_for(['food', 'lunch', 'dinner'], comment=True, comments=['Tasty!', 'Nice!', 'Yum!'])
session.end()# default enabled=False , enables the checking with the clarifai api (image# tagging) if secret and proj_id are not set, it will get the environment # variables 'CLARIFAI_API_KEY'session.set_use_clarifai(enabled=True, api_key='xxx')# uses the clarifai api to check if the image contains nsfw content# -> won't comment if image is nsfwsession.clarifai_check_img_for(['nsfw'])# checks the image for keywords food and lunch, if both are found,# comments with the given comments. If full_match is False (default), it only# requires a single tag to match Clarifai results.session.clarifai_check_img_for(['food', 'lunch'], comment=True, comments=['Tasty!', 'Yum!'], full_match=True)Check out https://clarifai.com/demo to see some of the available tags.
Use the nogui parameter to interact with virtual display
session = InstaPy(username='test', password='test', nogui=True)
Note: Chrome only! Must user chromedriver v2.9+
Use headless_browser parameter to run the bot via the CLI. Works great if running the scripts locally, or to deploy on a server. No GUI, less CPU intensive. Example
session = InstaPy(username='test', password='test', headless_browser=True)
Docker allows very easy and fast run of the instapy bot without any pain and tears.
Install docker from the official website https://www.docker.com/
Install VNC viewer if you do not have one. For windows, a good program is http://www.tightvnc.com/
Open docker_quickstart.py and fill the quotes after insta_username and insta_password with your credentials.
Don't forget to make other changes for the file as you want to. Read the documentation above for info.
First you need to open your terminal, move to the root folder (usually with the cd command) of instapy project and then type:
docker-compose up -d --buildThat's all! At this step, you are already successfully running your personal bot!
Run your VNC viewer, and type address and port localhost:5900. The password is secret.
Use your terminal again, type in the same window:
docker-compose downYour bot is stopped!
Those are just basic steps to run instapy bot on your PC with docker. There are other docker-compose settings file in the root of project.
Use it to help us with development and test instapy! docker-dev.yml file.
docker-compose -f docker-dev.yml up -dAfter striking this command, you can access your bot by VNC on the adress localhost:5901, the password is secret.
But there is more! There is a fully accessible bash console with all code mounted at the path /code. When you hack some files they are dynamically updated inside your container.
To access yor container console to run bot type localhost:22 in your favorite ssh client. The User is root and the password is root also.
Suitable to run in a remote server. Attention! You can not view what happened through VNC on this configuration docker-prod.yml file.
docker-compose -f docker-prod.yml up -dFirst you need to build the image by running this in the Terminal:
docker build -t instapy ./docker_conf/all_in_oneMake sure to use the nogui feature:
# you can use the nogui parameter to use a virtual displaysession=InstaPy(username='test', password='test', nogui=True)After the build succeeds, you can simply run the container with:
docker run --name=instapy -e INSTA_USER=<your-user> -e INSTA_PW=<your-pw> -d --rm instapyYou can add InstaPy to your crontab, so that the script will be executed regularly. This is especially useful for servers, but be sure not to break Instagrams follow and like limits.
# Edit or create a crontab
crontab -e
# Add information to execute your InstaPy regularly.
# With cd you navigate to your InstaPy folder, with the part after && # you execute your quickstart.py with python. Make sure that those paths match # your environment.
45 */4 * * * cd /home/user/InstaPy && /usr/bin/python ./quickstart.py
Automate with Schedule
Schedule is an in-process scheduler for periodic jobs that uses the builder pattern for configuration. Schedule lets you run Python functions periodically at pre-determined intervals using a simple, human-friendly syntax.
pip install schedulefrominstapyimportInstaPyimportscheduleimporttimedefjob():
try:
session=InstaPy(selenium_local_session=False) # Assuming running in Composesession.set_selenium_remote_session(selenium_url='http://selenium:4444/wd/hub')
session.login()
session.set_do_comment(enabled=True, percentage=20)
session.set_comments(['Well done!'])
session.set_do_follow(enabled=True, percentage=5, times=2)
session.like_by_tags(['love'], amount=100, media='Photo')
session.end()
except:
importtracebackprint(traceback.format_exc())
schedule.every().day.at("6:35").do(job)
schedule.every().day.at("16:22").do(job)
whileTrue:
schedule.run_pending()
time.sleep(1)Built-in delays prevent your account from getting banned. (Just make sure you don't like 1000s of post/day)
64-bit system is a requirement for current versions of chrome browser.