EmptyResponseException(result, "TikTok returned an empty response. They are detecting you're a bot, try some of these: headless=False, browser='webkit', consider using a proxy") TikTokApi.exceptions.EmptyResponseException: None -> TikTok returned an empty response. They are detecting you're a bot, try some of these: headless=False, browser='webkit', consider using a proxy

[Giaochan]

Describe the bug

TikTokApi.exceptions.EmptyResponseException is raised when iterating through video.comments(). TikTok returns an empty body, which the library interprets as the platform detecting automated traffic.

The buggy code

Smallest reproducible example with logging set to INFO:

py
Copiar
Editar
import asyncio
import logging
import os
from TikTokApi import TikTokApi

logging.basicConfig(level=logging.INFO) # <-- required for template
TikTokApi(logging_level=logging.INFO)

VIDEO_ID = 7275945937065741574
MS_TOKEN = os.getenv("MS_TOKEN") # redacted for security

async def get_comments():
async with TikTokApi() as api:
await api.create_sessions(
ms_tokens=[MS_TOKEN],
num_sessions=1,
sleep_after=3,
browser=os.getenv("TIKTOK_BROWSER", "chromium"),
)
video = api.video(id=VIDEO_ID)
async for comment in video.comments(count=30):
print(comment) # never reached – exception raised first

if name == "main":
asyncio.run(get_comments())
Expected behavior

The loop should print (up to) 30 comment objects for the specified video.

Error Trace (if any)

sql
Copiar
Editar
Traceback (most recent call last):
File "C:\Users\serro\Documents\tiktokapi\tiktokapifile.py", line 20, in
asyncio.run(get_comments())
...
File "...\TikTokApi\tiktok.py", line 503, in make_request
raise EmptyResponseException(result, "TikTok returned an empty response. They are detecting you're a bot, try some of these: headless=False, browser='webkit', consider using a proxy")
TikTokApi.exceptions.EmptyResponseException: None -> TikTok returned an empty response. They are detecting you're a bot, try some of these: headless=False, browser='webkit', consider using a proxy
Desktop (please complete the following information):

OS: Windows 10 Pro (22H2, build 19045)

Python: 3.11.6 (x64)

TikTokApi version: 5.6.2 ← (latest as of 2025-06-04)

pip show TikTokApi returns Version: 5.6.2

Browser: Playwright chromium (headless)

Additional context

Tried headless=False and browser="webkit" – same result.

ms_token freshly copied from a logged-in desktop Chrome session (~15 minutes old).

No proxy in use; traffic originates from a residential ISP IP in Ecuador.

Raising sleep_after to 10 s and lowering count to 5 did not help.

Happy to provide full INFO-level logs privately if needed.

Thanks for taking a look!

[Florararie]

Add headless=True as an argument to api.create_sessions and it should work. That's just unfortunately what it needs to work consistently.

[juancanosa17]

@Giaochan Has podido solucionarlo?

[Florararie]

Has podido solucionarlo?

Once again, just add headless=True as an argument to api.create_sessions. In my testing even with a proper msToken it will fail to pull the information correctly if you fail to set headless to True. I get that it's a pain but that's the only way it properly works unless you have a proxy setup.

It isnt a problem if you run it on a server as you're able to use xvfb-run before whatever process uses TikTokApi to run a non-headless browser without an x-server.

---

(Google Translated)

Nuevamente, simplemente agrega headless=True como argumento a api.create_sessions. En mis pruebas, incluso con un msToken correcto, no obtendrá la información correctamente si no configura headless como True. Entiendo que es un problema, pero es la única forma de que funcione correctamente a menos que tenga un proxy configurado.

No hay problema si lo ejecuta en un servidor, ya que puede usar xvfb-run antes de que cualquier proceso use TikTokApi para ejecutar un navegador sin interfaz gráfica sin un servidor x.