Twitter GraphQL API implementation with SNScrape data models.

pip install twscrape

Or development version:

pip install git+https://github.com/vladkens/twscrape.git

• Support both Search & GraphQL Twitter API
• Async/Await functions (can run multiple scrapers in parallel at the same time)
• Login flow (with receiving verification code from email)
• Saving/restoring account sessions
• Raw Twitter API responses & SNScrape models
• Automatic account switching to smooth Twitter API rate limits

This project requires authorized X/Twitter accounts to work with the API. You have two options:

1. Create Your Own Account: While you can register a new account on X/Twitter yourself, it's can be difficult due to strict verification processes and high ban rates.

2. Use Ready Accounts: For immediate access, you can get ready-to-use accounts with cookies from our recommended provider. Cookie-based accounts typically have fewer login issues.

For optimal performance and to avoid IP-based restrictions, we also recommend using proxies from our provider.

Disclaimer: While X/Twitter's Terms of Service discourage using multiple accounts, this is a common practice for data collection and research purposes. Use responsibly and at your own discretion.

import asyncio
from twscrape import API, gather
from twscrape.logger import set_log_level

async def main():
    api = API()  # or API("path-to.db") – default is `accounts.db`

    # ADD ACCOUNTS (for CLI usage see next readme section)

    # Option 1. Adding account with cookies (more stable)
    cookies = "abc=12; ct0=xyz"  # or '{"abc": "12", "ct0": "xyz"}'
    await api.pool.add_account("user3", "pass3", "[email protected]", "mail_pass3", cookies=cookies)

    # Option2. Adding account with login / password (less stable)
    # email login / password required to receive the verification code via IMAP protocol
    # (not all email providers are supported, e.g. ProtonMail)
    await api.pool.add_account("user1", "pass1", "[email protected]", "mail_pass1")
    await api.pool.add_account("user2", "pass2", "[email protected]", "mail_pass2")
    await api.pool.login_all() # try to login to receive account cookies

    # API USAGE

    # search (latest tab)
    await gather(api.search("elon musk", limit=20))  # list[Tweet]
    # change search tab (product), can be: Top, Latest (default), Media
    await gather(api.search("elon musk", limit=20, kv={"product": "Top"}))

    # tweet info
    tweet_id = 20
    await api.tweet_details(tweet_id)  # Tweet
    await gather(api.retweeters(tweet_id, limit=20))  # list[User]

    # Note: this method have small pagination from X side, like 5 tweets per query
    await gather(api.tweet_replies(tweet_id, limit=20))  # list[Tweet]

    # get user by login
    user_login = "xdevelopers"
    await api.user_by_login(user_login)  # User

    # user info
    user_id = 2244994945
    await api.user_by_id(user_id)  # User
    await gather(api.following(user_id, limit=20))  # list[User]
    await gather(api.followers(user_id, limit=20))  # list[User]
    await gather(api.verified_followers(user_id, limit=20))  # list[User]
    await gather(api.subscriptions(user_id, limit=20))  # list[User]
    await gather(api.user_tweets(user_id, limit=20))  # list[Tweet]
    await gather(api.user_tweets_and_replies(user_id, limit=20))  # list[Tweet]
    await gather(api.user_media(user_id, limit=20))  # list[Tweet]

    # list info
    await gather(api.list_timeline(list_id=123456789))

    # trends
    await gather(api.trends("news"))  # list[Trend]
    await gather(api.trends("sport"))  # list[Trend]
    await gather(api.trends("VGltZWxpbmU6DAC2CwABAAAACHRyZW5kaW5nAAA"))  # list[Trend]

    # NOTE 1: gather is a helper function to receive all data as list, FOR can be used as well:
    async for tweet in api.search("elon musk"):
        print(tweet.id, tweet.user.username, tweet.rawContent)  # tweet is `Tweet` object

    # NOTE 2: all methods have `raw` version (returns `httpx.Response` object):
    async for rep in api.search_raw("elon musk"):
        print(rep.status_code, rep.json())  # rep is `httpx.Response` object

    # change log level, default info
    set_log_level("DEBUG")

    # Tweet & User model can be converted to regular dict or json, e.g.:
    doc = await api.user_by_id(user_id)  # User
    doc.dict()  # -> python dict
    doc.json()  # -> json string

if __name__ == "__main__":
    asyncio.run(main())

In order to correctly release an account in case of break in loop, a special syntax must be used. Otherwise, Python's events loop will release lock on the account sometime in the future. See explanation here.

from contextlib import aclosing

async with aclosing(api.search("elon musk")) as gen:
    async for tweet in gen:
        if tweet.id < 200:
            break

# show all commands
twscrape

# help on specific comand
twscrape search --help

To add accounts use add_accounts command. Command syntax is:

twscrape add_accounts <file_path> <line_format>

Where: <line_format> is format of line if accounts file splited by delimeter. Possible tokens:

• username – required
• password – required
• email – required
• email_password – to receive email code (you can use --manual mode to get code)
• cookies – can be any parsable format (string, json, base64 string, etc)
• _ – skip column from parse

Tokens should be splited by delimeter, usually ":" used.

Example:

I have account files named order-12345.txt with format:

username:password:email:email password:user_agent:cookies

Command to add accounts will be (user_agent column skiped with _):

twscrape add_accounts ./order-12345.txt username:password:email:email_password:_:cookies

Note: If you added accounts with cookies, login not required.

Run:

twscrape login_accounts

twscrape will start login flow for each new account. If X will ask to verify email and you provided email_password in add_account, then twscrape will try to receive verification code by IMAP protocol. After success login account cookies will be saved to db file for future use.

In case your email provider not support IMAP protocol (ProtonMail, Tutanota, etc) or IMAP is disabled in settings, you can enter email verification code manually. To do this run login command with --manual flag.

Example:

twscrape login_accounts --manual
twscrape relogin user1 user2 --manual
twscrape relogin_failed --manual

twscrape accounts

# Output:
# username  logged_in  active  last_used            total_req  error_msg
# user1     True       True    2023-05-20 03:20:40  100        None
# user2     True       True    2023-05-20 03:25:45  120        None
# user3     False      False   None                 120        Login error

It is possible to re-login specific accounts:

twscrape relogin user1 user2

Or retry login for all failed logins:

twscrape relogin_failed

Useful if using a different set of accounts for different actions

twscrape --db test-accounts.db <command>

twscrape search "QUERY" --limit=20
twscrape tweet_details TWEET_ID
twscrape tweet_replies TWEET_ID --limit=20
twscrape retweeters TWEET_ID --limit=20
twscrape user_by_id USER_ID
twscrape user_by_login USERNAME
twscrape user_media USER_ID --limit=20
twscrape following USER_ID --limit=20
twscrape followers USER_ID --limit=20
twscrape verified_followers USER_ID --limit=20
twscrape subscriptions USER_ID --limit=20
twscrape user_tweets USER_ID --limit=20
twscrape user_tweets_and_replies USER_ID --limit=20
twscrape trends sport

The default output is in the console (stdout), one document per line. So it can be redirected to the file.

twscrape search "elon mask lang:es" --limit=20 > data.txt

By default, parsed data is returned. The original tweet responses can be retrieved with --raw flag.

twscrape search "elon mask lang:es" --limit=20 --raw

X API works through pagination, each API method can have different defaults for per page parameter (and this parameter can't be changed by caller). So limit param in twscrape is the desired number of objects (tweets or users, depending on the method). twscrape tries to return NO LESS objects than requested. If the X API returns less or more objects, twscrape will return whatever X gives.

There are few options to use proxies.

1. You can add proxy per account

proxy = "http://login:[email protected]:8080"
await api.pool.add_account("user4", "pass4", "[email protected]", "mail_pass4", proxy=proxy)

2. You can use global proxy for all accounts

proxy = "http://login:[email protected]:8080"
api = API(proxy=proxy)
doc = await api.user_by_login("elonmusk")

3. Use can set proxy with environemt variable TWS_RPOXY:

TWS_PROXY=socks5://user:[email protected]:1080 twscrape user_by_login elonmusk

4. You can change proxy any time like:

api.proxy = "socks5://user:[email protected]:1080"
doc = await api.user_by_login("elonmusk")  # new proxy will be used
api.proxy = None
doc = await api.user_by_login("elonmusk")  # no proxy used

5. Proxy priorities

• api.proxy have top priority
• env.proxy will be used if api.proxy is None
• acc.proxy have lowest priotity

So if you want to use proxy PER ACCOUNT, do NOT override proxy with env variable or by passing proxy param to API.

Note: If proxy not working, exception will be raised from API class.

• TWS_PROXY - global proxy for all accounts (e.g. socks5://user:[email protected]:1080)
• TWS_WAIT_EMAIL_CODE - timeout for email verification code during login (default: 30, in seconds)
• TWS_RAISE_WHEN_NO_ACCOUNT - raise NoAccountError exception when no available accounts, instead of waiting (default: false, values: false/0/true/1)

X/Twitter regularly updates their rate limits. Current basic behavior:

• Request limits reset every 15 minutes for each endpoint individually
• Each account has separate limits for different operations (search, profile views, etc.)

API data limitations:

• user_tweets & user_tweets_and_replies - can return ~3200 tweets maximum
• Rate limits may vary based on account age and status

• How to still scrape millions of tweets in 2023
• (Add Article)

• twitter-advanced-search – guide on search filters
• TweeterPy - Another X client
• twitter-api-client – Implementation of Twitter's v1, v2, and GraphQL APIs