Skip to main content

Getting Started

This guide gets PRISM running in under five minutes.

Prerequisites​

  • Docker (recommended) or a Linux server with a Rust toolchain
  • Google Chrome or Chromium installed and available in $PATH (the Docker image bundles Chromium automatically)

Quick Start with Docker​

docker run -d -p 4000:4000 -p 4001:4001 \
  -e RUST_LOG=info \
  ghcr.io/trident-cache/prism:1.0.0

Port 4000 serves proxied traffic. Port 4001 exposes the admin API.

By default PRISM expects your SPA origin at http://localhost:3000. Override it with a config file or environment variable (see Configuration).

Quick Start from Binary​

Download the latest release from GitHub Releases, then run:

chmod +x prism
./prism --config config.toml

Minimal Configuration​

Create a config.toml:

[server]
address = "127.0.0.1:4000"
origin  = "http://my-app:3000"
mode    = "bot-only"             # or "render-all"
  • address -- the host and port PRISM listens on.
  • origin -- the upstream SPA that PRISM will fetch pages from and render.
  • mode -- bot-only renders only for detected bots (default). render-all renders every request through headless Chrome.

Test It​

Send a request with a regular browser User-Agent:

curl -s -o /dev/null -w "%{http_code}" http://localhost:4000/
# 200 -- proxied directly from origin, no rendering

Now send a request as Googlebot:

curl -s -H "User-Agent: Googlebot" http://localhost:4000/ | head -20
# Fully rendered HTML with all JS-generated content

The X-Prism-Cache response header tells you whether the result was a cache hit or miss.

Verify with the Admin API​

Check that PRISM is healthy:

curl http://localhost:4001/health
# {"status":"ok"}

View Prometheus metrics:

curl http://localhost:4001/metrics

Next Steps​

  • Configuration -- full reference for config.toml options
  • Rendering -- tune Chrome pool size, timeouts, and wait strategies
  • Deployment -- production deployment patterns with Docker, systemd, and Kubernetes