Proxies for web scraping

A proxy is an exit. A scraper is a collector. I will not sell them as the same product. If you only need a different IP for a job you are allowed to do, an HTTP or SOCKS5 line may be enough. If you need a headed cookie store, that is a browser profile, and it is still not a license to hide a collector.

The useful caveat: MaskWright 0.1 is a local headed launcher. It is not a scraping browser. No Playwright product. We will not hide a scraper. We will not hide a scraper. Scraping ethics we follow.

When a proxy is enough

Official APIs that already accept a static or rotating exit. Your own servers. Jobs where robots.txt and terms already allow automated collection, and you are not trying to look like a shopper. What is web scraping. Web scraping with Python overview. Web scraping tools criteria.

MaskWright parses http, https (treated as an HTTP forwarder; TLS-to-proxy is not implemented), and socks5. We do not broker traffic. Provider templates exist for convenience. Credentials stay in the encrypted profiles.json index.

When a headed profile is the wrong tool

Sites that asked for a person. Inboxes. Anything you would have to disguise. LinkedIn we will not scrape inboxes. Instagram public looking versus collection: Instagram public research versus scrape.

A headed window with a proxy is still automation if a script drives it. What is a scraping browser. I do not document attaching drivers to MaskWright's internal debugging port.

Sticky versus rotating

Logins want sticky exits. Bulk collection, where it is allowed, often wants rotation. Do not put a rotating exit under a live store or ads login you still need tomorrow.

Geo: if you attach a proxy to a MaskWright profile, timezone and language can follow exit geo on launch. That is for operator profiles, not a scrape costume.

What I attach where

For an allowed collector I control, the proxy lives in the collector's config, not in a MaskWright profile I also use for Seller Central. Mixing those is how a rotate kills a login.

For headed public reading, I may attach HTTP or SOCKS5 to an empty MaskWright profile. Geo check. I still do not hide a loop. I read. I note. I stop at login walls I am not invited through.

HTTPS proxy type is parsed but TLS-to-proxy is not implemented. If a provider sold you that, it is their problem, not a MaskWright feature. SOCKS5 auth is in the handshake. HTTP auth uses a per-launch MV3 helper. Connectivity check is TCP only, so I still verify the first request.

Sticky with any login I must keep. Rotate only on jobs that are allowed to be anonymous collection. Do not put a mobile ASN in front of a desktop Client Hints profile and call it a phone.

Inboxes, Instagram private graphs, LinkedIn Recruiter dumps: stop. Official APIs or a person in a licensed seat.

FAQ

Do I need residential proxies to scrape?

Only if the job is allowed and the destination actually requires that class of exit. I do not treat residential as a magic bypass. MaskWright does not sell exits.

Can I use MaskWright as the scraping browser?

No. Use a collector you understand, against targets you are allowed to collect, with an exit you brought. Use MaskWright to sit in a chair on authorized logins.

Is a cloud scraping browser better?

Someone else holds the session and the HTML. That is a custody choice. Compare local versus cloud is about operator profiles, not a scrape endorsement.

Captcha farms in front of the proxy?

No.

Can I point MaskWright at a rotating pool and page through search?

That is a collector in a headed costume. If the site forbade it, stop. If an official API exists, use the API. Do not hide the loop in a profile you also use for logins.

Bring an exit for work you are allowed to do. Do not dress a collector as a shopper. Keep headed profiles for people.

I skip captcha farms, TLS-to-proxy claims we did not implement, and attaching drivers to the internal debug port. Sticky with logins. Rotate only on allowed anonymous collection in a tool that is actually a collector.

More notes sit under Scraping. Download MaskWright if you wanted a local operator profile instead of a scraper.