Proxy Formats and Parsers

Supported proxy formats, parser profiles, JSON paths, HTML selectors, and dedupe behavior.

ZeroTrace Proxy accepts a broad set of proxy input formats across extractor, checker, benchmark, leak test, chain tester, and viewer workflows.

Supported Proxy Formats

Format	Example
Host and port	`192.0.2.10:8080`
Scheme URL	`http://192.0.2.10:8080`
SOCKS URL	`socks5://198.51.100.20:1080`
User/pass URL	`http://user:pass@192.0.2.10:8080`
Host-port-user-pass	`192.0.2.10:8080:user:pass`
IPv6 bracketed	`http://[2001:db8::10]:8080`

Supported schemes include:

HTTP
HTTPS
SOCKS4
SOCKS4a
SOCKS5
SOCKS5h

Scheme variants with missing or malformed :// are normalized when possible.

Endpoint Validation

The parser validates:

IPv4 hosts
IPv6 hosts
domains
localhost
ports from 1 to 65535

Parser Profiles

Profile	Behavior
Auto	Chooses JSON, HTML, or plain text based on content type, body shape, selector, and JSON path.
Plain text	Runs text extraction directly.
HTML	Uses selectors, common elements, attributes, tables, lists, and script-like blocks.
JSON	Walks JSON payloads and proxy-shaped objects.

JSON Path Examples

data.items[*]
data.items[*].proxy
sources[0].items[*]

Supported path features:

dot paths
numeric indexes like [0]
wildcards like [*]

HTML Selector Examples

table tbody tr
pre
code
[data-proxy]

Use selectors when Auto mode finds too much or too little.

Extractor Sources and Fetch Controls

Feature	Detail
Sources	Single source, many sources, and local file.
HTTP controls	Timeout seconds, concurrency, redirects, random User-Agent, custom User-Agent, custom headers, and cookies.
Body handling	Gzip, zlib, and deflate-style decompression plus an extraction body cap around 4 MiB.
Per-source metrics	HTTP status, duration, content type, response bytes, parser used, extraction scope count, candidate count, duplicate count, proxy count, and error.

HTML mode can read selected elements, table tr, pre, code, textarea, ordered lists, unordered lists, data-ip, data-host, data-proxy, and data-address. It also reads useful attributes such as data-*, href, content, value, and title, understands table cells and child elements, and inspects embedded scripts that look like JSON or contain proxy-shaped values.

JSON Extraction Details

JSON mode collects strings, numbers, arrays, and object fields. It can infer proxy-shaped objects from host-like fields, port fields, scheme/protocol/type fields, username/user/login fields, and password/pass fields.

Dedupe Modes

Mode	Best for
Full	Keeping scheme/auth variants separate.
Host + port	Reducing duplicates where only scheme or auth varies.
Host	Aggressive cleanup by host.

For mixed public proxy sources, start with Auto parser and Host + port dedupe. Tighten to HTML selector or JSON path only when source metrics show noise.