Failed to parse scripts

[smmrokn]

2025-12-04 08:14:42.551 | WARNING | twscrape.queue_client:req:291 - Unknown error. Account timeouted for 15 minutes. Create issue please: https://github.com/vladkens/twscrape/issues If it mistake, you can unlock accounts with twscrape reset_locks. Err: <class 'Exception'>: Failed to parse scripts

[kunalransing]

@vladkens We have been using it for quite some time, but this error has started occurring suddenly. Even after running reset_locks, the issue remains.
Any idea ? Thanks.

[JamesMcMinn]

See #280

I'm not sure why the above issue was closed. I've proposed a fix in the issue, however the actual library still needs to implement it.

The tldr; is that Twitter is returning some invalid JSON, so the fix is to try and fix the JSON before parsing.

[kunalransing]

@JamesMcMinn Hey, Thanks. I added that code. Now its working fine.

[Tormented]

i'm experiencing same issue. is it fixed already? i'm guessing it's cus of api change?

[kunalransing]

i'm experiencing same issue. is it fixed already? i'm guessing it's cus of api change?

I guess its not fixed on repo. You need to manually fix on file directly xx/venv/lib/python3.13/site-packages/twscrape/xclid.py
Check code mentioned by @JamesMcMinn in #280

[michael-bey]

This just happened to me too, was working fine up until today

[xxchen04120412]

Still hitting this on 0.17.0. The issue is that x.com's script map contains unquoted JS identifier keys (e.g. node_modules_pnpm_...) which are valid JS but not valid JSON, so json.loads() fails.

Fix — add a regex fallback to xclid.py::get_scripts_list():

def get_scripts_list(text: str):
scripts = text.split('e=>e+"."+')[1].split('[e]+"a.js"')[0]
try:
for k, v in json.loads(scripts).items():
yield script_url(k, f"{v}a")
except json.decoder.JSONDecodeError:
fixed = re.sub(r'([,{])\s*(\w+)\s*:', r'\1"\2":', scripts)
for k, v in json.loads(fixed).items():
yield script_url(k, f"{v}a")
Tested and working on 0.17.0 with cookie-based auth (auth_token + ct0).

[michael-bey]

Still hitting this on 0.17.0. The issue is that x.com's script map contains unquoted JS identifier keys (e.g. node_modules_pnpm_...) which are valid JS but not valid JSON, so json.loads() fails.

Fix — add a regex fallback to xclid.py::get_scripts_list():

def get_scripts_list(text: str): scripts = text.split('e=>e+"."+')[1].split('[e]+"a.js"')[0] try: for k, v in json.loads(scripts).items(): yield script_url(k, f"{v}a") except json.decoder.JSONDecodeError: fixed = re.sub(r'([,{])\s*(\w+)\s*:', r'\1"\2":', scripts) for k, v in json.loads(fixed).items(): yield script_url(k, f"{v}a") Tested and working on 0.17.0 with cookie-based auth (auth_token + ct0).

Can you try my fix please? #288