SpotwebNG

A modern Spotweb. Self-hosted Spotnet indexer for Usenet, with a Newznab API for Radarr, Sonarr and Prowlarr — and a UI you'll actually want to browse.

What it is

Spotnet is a Usenet community that publishes signed "spots" — posts describing content available on Usenet. Spotweb has indexed them since 2011 and still works, but it is PHP from another era.

SpotwebNG is a ground-up rebuild in Python: the same Spotnet retrieval engine and category system, a Newznab API your existing tools already speak, and a browsing experience built for how people actually look for things.

It is not a fork. The wire format, category tree and filter grammar were re-derived from the Spotweb source so behaviour matches; everything else is new.

Why you might want it

	Spotweb	SpotwebNG
Stack	PHP 7, jQuery	Python 3.11, FastAPI, async SQLAlchemy
Install	web server + PHP + database	docker compose up -d
Browsing	dense tables	poster grid with cover art, or a table if you prefer
Filtering	category tree	the same tree grammar, with live counts on every option
Accounts	full ACL system	admin / viewer roles, throttled login
API	Newznab	Newznab, per-client keys with last-used tracking

Screenshots

Sub-filters are contextual: Movies get Format, Source, Language and Genre; Games get Platform, Format and Genre. Every option shows how many spots it would leave.

Run it

git clone https://github.com/FornaxMedia/SpotwebNG.git
cd SpotwebNG
docker compose up -d

Open http://localhost:7979. The setup wizard asks for an administrator login first — complete it, because until you do the instance is open to anyone who can reach it — then your Usenet provider, picked from a built-in catalogue of 68 with host, port and TLS filled in for you. The connection is tested before anything is saved.

Everything lives in one Docker volume: the database, cached cover art and assembled NZBs.

Without Docker

python -m venv .venv && .venv/bin/pip install -r requirements.txt
.venv/bin/uvicorn app.main:app --host 0.0.0.0 --port 7979

SQLite is the zero-config default; set SPOTWEBNEXT_DATABASE_URL to a postgresql+asyncpg:// URL to use Postgres instead.

Connect your download tools

SpotwebNG speaks the Newznab API, which almost every Usenet tool already understands. Get a key from Settings → API access (one per client, so you can revoke one without breaking the rest), then point tools at:

http://your-host:7979/api

Worth knowing before you start: SABnzbd and NZBGet are downloaders, not searchers. They fetch NZB files but do not query indexers. The search side is handled by Prowlarr or the *arr apps, which then hand the NZB to your downloader.

Indexer managers — configure once, share everywhere

Tool	Where	What to enter
Prowlarr	Indexers → Add → Generic Newznab	URL http://your-host:7979, API path /api, your API key
NZBHydra2	Config → Indexers → Add → Newznab	Host http://your-host:7979, API key, then Check caps

This is the recommended route: add SpotwebNG once in Prowlarr and it syncs to Radarr, Sonarr, Lidarr and Readarr automatically.

The *arr apps — add it directly

In Settings → Indexers → Add → Newznab:

App	Categories to select	Search it uses
Radarr	2000 Movies	t=movie with imdbid or a title
Sonarr	5000 TV	t=tvsearch with season and episode
Lidarr	3000 Music	t=music with artist and album
Readarr	7000 Books	t=book with author and title
Whisparr	6000 Erotica	t=search — see the note below

Leave Additional Parameters empty. If a client reports "indexer does not support this search", check that its category is present in http://your-host:7979/api?t=caps.

Adult content is hidden by default. While Settings → Display → Hide adult content is on, the 6000 range is not advertised in caps, is not returned by any search, and its covers are not served — so Whisparr will find nothing until an administrator turns the setting off. Only administrators can change it.

Mylar3 asks for comics (category 7030), which Spotnet does not file separately, so it has nothing useful to match against.

Downloaders — subscribe to a feed

The API returns a standard RSS 2.0 feed whose items carry an NZB <enclosure>, so a downloader can grab new spots without any indexer app in between.

SABnzbd — Config → RSS → add a feed:

http://your-host:7979/api?t=search&cat=2040&apikey=YOUR_KEY

Set a filter (for example Accept on a title pattern), tick Read the feed, and SABnzbd will download matching spots as they appear. Swap cat= for the categories you want; 2040 is Movies/HD, 5040 is TV/HD, 3040 lossless music.

NZBGet has no built-in RSS reader — use Prowlarr or an *arr app in front of it, or one of the community RSS extension scripts.

Mobile and browser

• nzb360 (Android) — add under Indexers as Newznab with the URL and key.
• LunaSea / nzbUnity — anywhere that accepts a Newznab indexer works.

Check it yourself

curl "http://your-host:7979/api?t=caps"                                   # no key needed
curl "http://your-host:7979/api?t=search&q=ubuntu&apikey=YOUR_KEY"        # a search
curl -O "http://your-host:7979/api?t=get&id=123&apikey=YOUR_KEY"          # a real NZB

If t=caps answers but a client refuses, it is nearly always the API path: some clients want the base URL, others the full /api.

How it works

• Retrieval — reads spot headers from the free.pt newsgroup over NNTP in XOVER batches, unpacking the metadata Spotnet packs into the From header. The position bookmark is validated by message-id, so article renumbering on the server cannot make the indexer skip or replay a range.
• Enrichment — fetches each spot's full XML payload for its description, cover art and NZB segment list. Covers are cached and served locally.
• Filtering — implements Spotweb's tree grammar (cat0_z0_a9, ~ to exclude): OR within a dimension, AND across dimensions. All 283 Spotnet category codes are included under their real names.
• Retention — keep spots listed while dropping their heavy parts, or delete them outright. Without a policy an index grows forever.

Configuration

Everything is configurable in the UI. Environment variables (prefix SPOTWEBNEXT_) exist for deployments:

Variable	Default	Purpose
DATABASE_URL	SQLite file	postgresql+asyncpg://… for Postgres
SEED_DEMO_DATA	true	demo spots on first boot so the UI is not empty
APP_NAME	SpotwebNG	shown in the UI and in Newznab caps

SPOTWEBNG_DATA sets where cover art and NZBs are written; the Docker image points it at the volume.

Status

Working today: retrieval, enrichment, filtering, the Newznab API, accounts and retention. Not yet: comments and reports groups, saved filters, per-user watchlists.

Contributions welcome. Tests run with pytest; ruff and mypy gate every commit.

Credits

The Spotnet wire format, category tree and filter semantics are derived from Spotweb by the Spotweb contributors. This project reimplements them and shares no code.

License

Apache 2.0 — see LICENSE.