User guides on Praxis

Run your first reverse proxy

Mon, 01 Jan 0001 00:00:00 +0000

Use Praxis to forward a request to a real HTTP backend. Configure its built-in router and load_balancer filters in YAML.

You need Docker Engine, curl, and an internet connection to pull the two images. The Praxis image listens on container port 8080; the echo service listens on port 3000.

Start the echo service

Create a Docker network so Praxis can reach the echo service by its container name:

docker network create praxis-quickstart
docker run --detach --name praxis-echo \
 --network praxis-quickstart \
 registry.k8s.io/gateway-api/conformance/echo-basic:v0.1.0

The echo service listens on port 3000 inside the network. It returns JSON describing the request it received.

Install and run Praxis

Mon, 01 Jan 0001 00:00:00 +0000

Praxis v0.7.2 is the current default documentation release. Use its image for built-in filters and ordinary proxy tasks. Build from source only when developing Praxis or enabling a feature that requires a custom build.

Choose how to run it

Container image

Pull the version matching your documentation:

docker pull ghcr.io/praxis-proxy/praxis:0.7.2

For the FIPS variant of the same release, append -fips to the tag:

docker pull ghcr.io/praxis-proxy/praxis:0.7.2-fips

The Praxis image package lists supported tags and platforms. For FIPS, use the same release tag with -fips appended. Praxis AI has its own release numbers; its current documentation release is v0.8.0. See the Praxis AI image package for its standard and FIPS image tags.

Choose and configure filters

Mon, 01 Jan 0001 00:00:00 +0000

Praxis applies filters to requests as they pass through a proxy. A filter is a named rule that can forward a request, change a header, or return a response. You choose and configure built-in filters in a YAML file; this does not require writing Rust.

Configure built-in filters in YAML; no Rust code is needed. If you have not run Praxis yet, start with the first reverse-proxy tutorial. Use the developer path for custom filters or features that require a source build.