snscrape not working
snscrape's Twitter module has failed for every user since 30 June 2023, and its maintainer says it cannot scrape Twitter anymore. Most other failures come from Instagram, or from a Python release newer than any snscrape was built for.
Three checks first
1. Which module you run. Whether a fix exists at all depends on the module. The maintainer's latest status for each, posted in issue #1045 on 13 February 2025:
| Module | The maintainer's words |
|---|---|
| Telegram, Weibo | "Telegram and Weibo work." |
| Mastodon | "Mastodon should work on the few 3.x instances that might still exist. Broken on 4.x instances." |
| Facebook, Instagram | "Facebook and Instagram have had issues for a long time due to restrictions and brokenness on their side." |
| "Reddit depended on Pushshift, which is no longer open to the public." | |
| "Twitter is broken due to the login wall." | |
| VKontakte | "VK is broken due to site changes." |
2. Your Python version. On Python 3.12 or newer, snscrape stops while it imports its modules, before a single request leaves your machine:
AttributeError: 'FileFinder' object has no attribute 'find_module'
The loader in snscrape/modules/__init__.py imports each module with this line, and Python 3.12 removed the find_module method it relies on:
module = importer.find_module(moduleName).load_module(moduleName)
The release names Python 3.8 to 3.11 in its classifiers, so give it a 3.11 environment of its own:
python3.11 -m venv snscrape-env snscrape-env/bin/pip install snscrape snscrape-env/bin/snscrape --version
3. Your snscrape version, and why upgrading has stopped helping.
snscrape --version
0.7.0.20230622 is the newest release on PyPI, and the repository's last commit dates from that same day, 22 June 2023. During earlier breakages the maintainer put fixes in the development version first ("The fix is currently only available in the development version", #846), but there is nothing newer there now: installing from the repository gets you the same code as pip install snscrape.
Every message on these pages is copied from snscrape 0.7.0.20230622.
Find your message
| Message or log line | What happened | Page |
|---|---|---|
AttributeError: 'FileFinder' object has no attribute 'find_module' | Python 3.12 or newer | Check 2 above |
Error retrieving ...: blocked (404) | X refused the request, as it has every Twitter scrape since June 2023 | Blocked |
blocked (403) or blocked (429) | A refusal or a rate limit on the guest token | Blocked |
Unable to retrieve guest token | X gave no guest token | Guest token |
... guest/activate.json: non-200 response (429) | Too many guest tokens asked for | Guest token |
4 requests to ... failed, giving up. | Every attempt at one address failed; the Errors: line under it says why | Giving up |
EntityUnavailable: User unavailable or Community unavailable | X marked the account or community as one it will not show | User unavailable |
ScraperException: Empty response | X sent back no account or community at all | Empty response |
Redirected to login page, or IndexError: list index out of range on instagram-* | Instagram sent the scraper to its sign-in screen | Instagram login |
Is the Twitter module coming back
On 30 June 2023 the maintainer opened issue #996, which is still pinned and open: "With the exception of twitter-trends, all Twitter scrapes are failing since sometime in the past hour." They labelled it bug, module:twitter and upstream.
In issue #1037, on 23 February 2024, they went further: "snscrape can't scrape Twitter anymore." The most it might do in future, they wrote, is fetch individual tweets and "a profile's most popular tweets", and "I will likely remove all other Twitter scrapers." Accounts are ruled out as a way back: "Anything behind authentication walls has always been outside of snscrape's scope and design goal." No commit has landed since.
Where each snscrape job can go
SocialScrape covers several of the jobs snscrape's broken modules did, with an API key in place of a scraper to keep alive. Not every module has a match, and the matches differ from snscrape in ways worth checking first:
| snscrape scraper | SocialScrape endpoint | How it differs |
|---|---|---|
twitter-user, twitter-profile | GET /v1/twitter/profile, GET /v1/twitter/profile/tweets | About 20 popular posts and reposts, and no older pages |
twitter-tweet | GET /v1/twitter/tweet | One post with its counts and media |
twitter-search, twitter-hashtag | GET /v1/twitter/search/posts | Posts a web search engine has indexed on x.com, not X's own search |
twitter-community | GET /v1/twitter/community/tweets | The 20 most liked posts, which is all X shows |
instagram-user, instagram-hashtag | GET /v1/instagram/profile/posts, GET /v1/instagram/search/hashtag | Posts 12 a page; hashtag results are best effort |
reddit-subreddit, reddit-search | GET /v1/reddit/subreddit/posts, GET /v1/reddit/search | Reddit's own search, posts only, with no Pushshift archive behind it |
telegram-channel | GET /v1/telegram/channel/posts | snscrape's Telegram module still works, so this one is a choice rather than a rescue |
twitter-trends, mastodon-*, vkontakte-user, weibo-user | Not offered |
Each of these endpoints is 1 credit when it returns data, and an error charges nothing. Sign-up comes with 100 credits, enough to rerun the accounts your snscrape jobs used to cover.
Questions, answered.
Common questions
Is snscrape still maintained?
The repository's last commit is from 22 June 2023. In November 2023 the maintainer wrote in #1037: "I do intend to resume development, but I can't currently say when that will happen."
Will snscrape add a login for Twitter or Instagram?
No. The maintainer has said so for both: authentication sits outside the project's scope (#1037), and for Instagram, "I have no intentions of adding any sort of login to snscrape" (#165).
Does installing snscrape from GitHub fix it?
Not anymore. The repository and the PyPI release stop on the same date, so either route installs the code of 0.7.0.20230622.
Which snscrape modules still work?
Telegram and Weibo, by the maintainer's account in February 2025, and Mastodon on the few servers still running version 3.
Start pulling social data.
100 free credits. No card needed.