nostream: your relay, your rules

/hero · nostream 3.5.0

01 Who it's for

One relay. Three ways to run it.

Whether you are opening a relay to everyone, charging for access, or keeping it private, nostream ships the knobs you need, without forking.

PUB 01

Public relay operators

Serve the open network without getting flattened by it.

Sliding-window rate limits with backoff hints clients can respect
Adaptive proof-of-work, event size and created_at limits
Horizontal workers with Redis fan-out

PAID 02

Paid relay operators

Turn a relay into a service. Admission fees settle over Lightning.

ZEBEDEE, Nodeless, OpenNode, LNbits and LNURL processors
Admission checks and NIP-43 invite codes
Admission fee schedules with pubkey whitelists

SELF 03

Self-hosters

A relay for you, your family or your community, on your own box.

One command: nostream start, with Tor, I2P or nginx + certbot
Import and export events as JSON or JSONL
Restricted kinds, pubkey allow lists and NIP-42 auth

02 Feature pillars

Everything a relay needs. Nothing it doesn't.

Protocol breadth, network reach and the ops tooling to keep it healthy, in one codebase you can read.

NIP

Protocol breadth

25 NIPs implemented relay-side, from basic flow to auth, counts, search and encrypted payloads, plus Marmot MIPs 0–3 for E2EE group messaging.

NET

Tor & I2P built in

Publish an .onion or .i2p address with a flag. nostream start --tor --i2p, then nostream info --tor-hostname.

OPS

Admin console

Live metrics, NIP-66 network health, NIP-11 mismatch warnings and a validated settings editor.

Paid admission

Charge to publish. Invoices via ZEBEDEE, Nodeless, OpenNode, LNbits or any LNURL-verify provider.

LIM

Limits that hold

Rate limits, adaptive proof-of-work, kind and size rules. Clients get a rate-limited hint with a backoff.

I/O

Your data, portable

Import and export events as JSON or JSONL in batches. Back up, migrate, or seed a new relay.

$ nostream export --output backup.jsonl --format jsonl

03 Admin & observability

Know what your relay is actually doing.

A built-in console for operators: no SSH, no SQL, no guessing at 3am.

Login-gated with IP allow lists and rate limits
Prometheus-compatible metrics and OpenTelemetry hooks
/readyz for load balancers and blue/green deploys
Self-hosted probes, labelled as one monitor among many

nostream admin console: Live metrics (1) — Live metrics

Events per second, rejections, connections, CPU and memory, streamed over SSE.

nostream admin console: Network health — Network health

NIP-66 probes for DNS, TLS, WebSocket read/write RTT and NIP-11, per target.

nostream admin console: NIP-11 mismatch warnings — NIP-11 mismatch warnings

Flags when what your relay advertises disagrees with what probes measure.

nostream admin console: Settings editor — Settings editor

Guided, validated edits with automatic backups and one-click restore.

nostream admin console: Notifications — Notifications

Outbound alerts with a delivery log, so you hear about trouble first.

04 NIPs supported

25 NIPs, advertised in NIP-11.

The same list your relay returns as supported_nips, read straight from the repository at build time.

Marmot MIPs 00, 01, 02, 03: end-to-end encrypted group messaging over Nostr, advertised as supported_mips. →

Ship it

Run your own relay tonight.

Clone, set a secret, start. Docker Compose does the rest.

# clone and configure
$ git clone https://github.com/cameri/nostream.git && cd nostream
$ echo "SECRET=$(openssl rand -hex 128)" > .env
$ pnpm add -g . && nostream start

Choose a deploy path → Open an issue

nostream — a production-ready Nostr relay. Your relay, your rules.