About ZigProxy

ZigProxy is a technical reference for proxy engineering: forward proxies, reverse proxies, and the protocols, headers and configuration semantics that decide how a request travels from a client to an origin and back.

Who it is for#

Backend engineers, site reliability engineers, platform engineers and network engineers who are designing, operating or debugging a proxy layer. The site assumes you know what HTTP is, what TCP is, and how to read a configuration file. It does not assume you have memorised RFC 9110 or the difference between proxy_buffer_size and proxy_buffers.

What it covers#

Six clusters, and the boundaries between them are deliberate:

• Proxy fundamentals: the model, the vocabulary and the protocols underneath.
• Client IP and forwarding headers: X-Forwarded-For, RFC 7239 Forwarded, the PROXY protocol, and trust.
• Reverse proxy configuration: nginx, HAProxy, Envoy, Caddy and Traefik.
• TLS and proxy security: termination topologies, mTLS, interception, smuggling, SSRF.
• Performance and protocols: connection reuse, buffering, timeouts, HTTP/2 and HTTP/3, caching.
• Troubleshooting: reading proxy errors and bisecting a chain.

Plus six interactive tools that run entirely in the browser.

What it does not cover#

Deliberately out of scope, so you are not misled about what you will find here:

• Commercial proxy services. Residential, datacentre and mobile proxy providers, pricing comparisons and affiliate reviews. This site is about the technology, not the market.
• Web scraping strategy. Anti-bot evasion, fingerprint spoofing and CAPTCHA handling.
• Circumvention tooling. Guides to evading network filtering or geographic restrictions.
• Attack instructions. Security topics are covered from the perspective of building and auditing a defensible proxy layer. Pages explain vulnerability classes, how to detect them in your own systems, and how to fix them. They do not provide exploitation walkthroughs for systems you do not run.

The principle behind the content#

Vendor documentation describes one product accurately. It cannot describe the seams, and the seams are where proxy problems live: two correct implementations that disagree about the same input, a default chosen for a workload nobody has any more, a header that one hop appends and another overwrites.

So the recurring structure of a ZigProxy page is a comparison across implementations, a worked example with real configuration or real wire bytes, and the failure mode each choice produces, named by the log line you will actually see. If a page cannot offer at least one of those, it does not get published.

Independence#

ZigProxy has no commercial relationship with any of the software it covers. Nothing on the site is sponsored, and no vendor has been given review or approval over what is written about their product. Where a product is bad at something, the page says so, and where the open source edition of a product lacks a feature its commercial edition has, that is stated plainly rather than elided.

The site may display advertising. Advertising has no influence on which topics are covered, what is recommended, or how a product is assessed.

Getting in touch#

Corrections, version-specific behaviour we have missed, and requests for topics are all welcome. See contact, and see corrections for how changes after publication are logged.