Google blocks after 2captcha solve with residential proxies - TLS fingerprint issue?

Hi,

I'm running a Google CTR bot using residential proxies (Decodo - Turkey IPs) and 2captcha for recaptcha solving.

The bot uses Playwright with stealth. Proxies are residential. Captcha solves successfully, but immediately after solving, Google redirects back to "We detected unusual traffic" or the "I'm not a robot" page again. The token seems to be rejected.

What works:

- The bot searches Google and finds the target

- 2captcha returns a token

What fails:

- Google doesn't accept the token after solving

- We get stuck in a loop of solving and blocking

What we've tried:

- Different residential proxies (same result)

- Slowing down the bot (human-like delays)

- Different user-agents

- SOCS cookies injected

No luck.

I suspect the issue is TLS fingerprint mismatch between Playwright and real Chrome, or maybe the 2captcha solver is using a different IP than my proxy (IP mismatch).

Questions:

1. Has anyone solved this exact problem with Google?

2. Should I switch from Playwright to curl_cffi with impersonation?

3. Does 2captcha need to use the same proxy IP as my bot? If so, how to configure that?

4. Is there a ready-to-use solution or service that works reliably right now?

Any technical advice is appreciated. Thanks.