Running a Yakr relay on your homelab

A practical intro to operating a mailbox relay for your circle — pairing, ports, and what relays actually see.

Running a mailbox relay is one of the most concrete ways to participate in Yakr today. You are not opening a global messaging service — you are offering store-and-forward for people you have paired with.

What a relay does

A mailbox relay:

• Accepts POST /v1/blobs with opaque ciphertext and a 32-byte mailbox tag
• Stores blobs until expiry (typically ≤24h)
• Serves GET /v1/blobs/{tag} to authorised paired clients
• Validates protocol structure, rate limits, and tickets — but never decrypts message content

Typical homelab layout

Many operators run the reference relay on a high port (e.g. 8090) behind their existing reverse proxy or on Tailscale. The homelab relay guide in the repository covers deploy scripts and operator pairing.

The yakr.co.uk website is separate infrastructure — nginx serving static Astro output — and should not be confused with Yakr protocol relays.

Before you invite anyone

1. Pair with your relay identity (operator CLI flow)
2. Share your relay URL and TLS pin through a signed delivery profile after invite pairing
3. Understand experimental maturity — see SECURITY.md

Next steps

• Deploy with ./scripts/homelab_relay_deploy.sh (see repo docs)
• Read delivery & relays for how ciphertext routes through your node
• Join GitHub discussions for operator questions

← All posts