nostrfy

Introduction

What nostrfy is, what it does out of the box, and the ideas that shape how it is built and run.

nostrfy is a relay server for the Nostr protocol. It stores events (posts, reactions, profiles, …) sent by clients and delivers them in response to subscription requests.

Key features

  • Simple and stable — written in Rust; a single binary does everything.
  • Fast storage and search — LMDB database with a full-text search index (NIP-50).
  • Broad NIP support — 36 NIPs implemented, plus the Blossom file server: deletion, proof-of-work, delegation, groups, search and a management API.
  • Easy to operate — daemon mode, log rotation, hot configuration reload (SIGHUP), live statistics, a REST API and Prometheus metrics.

What you get out of the box

A single nostrfy binary on one port serves:

  • The WebSocket relay and the NIP-11 information document.
  • A read-only REST API at /api/v1/... — its own reader thread means REST traffic can never stall WebSocket subscribers.
  • The Blossom file server (media hosting) on its own hostname.
  • The NIP-86 management RPC, health check and Prometheus /metrics.

Resource profile

nostrfy is verified to run on a 0.25 vCPU / 512 MB VPS. The LMDB memory map is a sparse 1 TiB virtual reservation — physical disk grows only with real data — and process memory stays flat: a relay with a 252 MB database held 7.9 MB of private RSS.

A quick feel for it

Install, initialize and start in three commands:

sh
curl -fsSL https://raw.githubusercontent.com/iqbqioza/nostrfy/main/install.sh | sh
nostrfy --config nostrfy.toml init
nostrfy --config nostrfy.toml start

Then verify the relay is up:

sh
curl http://127.0.0.1:8080/health

And fetch its NIP-11 document:

sh
curl -H "Accept: application/nostr+json" http://127.0.0.1:8080/
Where to go next
Follow the quick start for your first run, or skip straight to the configuration reference to tune the relay.