tls-impersonate-proxy 0.1.0

Fetch URLs with a browser-matched TLS/JA3 fingerprint (via curl_cffi) through rotating authenticated proxies, with automatic anti-bot block detection and retry.

tls-impersonate-proxy

Fetch URLs with a real browser's TLS/JA3 fingerprint (via curl_cffi) through rotating authenticated proxies, with automatic anti-bot block detection and retry.

Why your scraper gets blocked before it even sends a request

Most "the proxies don't work" problems aren't an IP problem — they're a TLS problem.

Before a single byte of your HTTP request is parsed, the server sees how your client negotiates TLS: the cipher suites, extensions and curve order. It hashes that into a JA3/JA4 fingerprint. Python requests, httpx, aiohttp, curl and Go's net/http each produce a handshake no real Chrome or Safari ever sends — so you can be on a pristine residential IP, with perfect headers, and still get fingerprinted as a bot on the first packet.

tls-impersonate-proxy fixes the layer that rotating more proxies never touches:

• Browser-matched JA3 — borrows real Chrome / Safari / Edge TLS profiles from curl_cffi, so the handshake lines up with your User-Agent.
• Proxy rotation — round-robins a list of authenticated proxies (HTTP / SOCKS5).
• Block detection + retry — a 403/429/503 or a Cloudflare / DataDome / PerimeterX / Akamai challenge page triggers a real retry on a fresh exit instead of silently returning garbage.

Install

pip install tls-impersonate-proxy

CLI

# Fetch with a rotating browser fingerprint through a residential proxy
tls-fetch https://example.com -x socks5h://USERNAME:[e-mail]:913

# Pin a specific browser, print status + headers only
tls-fetch https://tls.browserleaks.com/json --impersonate chrome -i

# Rotate across several exits, retry up to 6 times on a block
tls-fetch https://target.site \
  -x socks5h://USERNAME:[e-mail]:913 \
  -x socks5h://USERNAME:[e-mail]:913 \
  --retries 6

Library

from tls_impersonate_proxy import ImpersonateClient, BlockedError

client = ImpersonateClient(
    proxies=[
        "socks5h://USERNAME:[e-mail]:913",
        "socks5h://USERNAME:[e-mail]:913",
    ],
    impersonate=None,   # None = rotate chrome/safari/edge fingerprints
    max_retries=4,
)

try:
    resp = client.get("https://target.site")
    print(resp.status_code, len(resp.text))
except BlockedError as e:
    print("still blocked:", e)

One-shot helper:

from tls_impersonate_proxy import fetch
html = fetch("https://target.site",
             proxies=["socks5h://USERNAME:[e-mail]:913"]).text

The fingerprint has to match the exit

A browser-matched JA3 gets you past the TLS check, but the exit IP still has to read as a real person. A datacenter ASN is scored as automation no matter how perfect the handshake is — the two signals have to agree.

• Verify the JA3 your client actually sends (and the exit ASN) with the free tool: check.jibaoproxy.com
• Residential exits with clean ASNs and sticky sessions: jibaoproxy.com — $5/GB residential, $0.8/GB datacenter, $5 free credit on signup.

Further reading

• JA3/TLS fingerprint detection, explained
• Bypassing DataDome & PerimeterX in 2026
• Headless browser detection & CDP leaks

License

MIT — see LICENSE. Use responsibly and respect target sites' terms and robots.txt.