WattsNew

A self-hosted, text-first news reader — built for the Tesla browser first, then refined for phone and desktop.

WattsNew started life on the big center screen of a Tesla Model Y: large type, generous tap targets, and a clean two-pane layout meant to be read while parked or charging. It works just as well on a phone or laptop. A built-in curator fetches fresh articles from RSS feeds across categories (AI, Tech, Science, Business, Gaming, Politics, Entertainment, Local…) on a schedule; the web reader lets you browse by category, read full articles in-place, and thumbs-up/down stories to tune what you see. It all runs in one small Docker container.

• 🚗 Tesla-first, everywhere-friendly — designed for the in-car browser (large type, big tap targets, two-pane reading), then refined for phones and desktop.
• 🔒 Fully self-hosted & private — no tracking, no third-party services. Runs on your own machine in a single Docker container.
• 🧠 No AI required — summaries come straight from the feeds. Optionally plug in Claude or OpenAI for nicer summaries.
• 📖 Read in-place — the reader extracts the full article (with images), so you rarely leave the app. Headless-browser fallback for stubborn sites.
• 👍 Personalized — your reactions boost sources you like and bury ones you don't, with per-category source diversity.
• ♿ Text-first & accessible — built for comfortable reading: large adjustable type, light/dark, keyboard-friendly.

Status: early but fully working. See the Roadmap.

Quick start

Requires Docker.

git clone https://github.com/wrdodd/wattsnew.git
cd wattsnew
cp .env.example .env          # set DASHBOARD_PASSWORD (and anything else)
docker compose up -d --build  # → http://localhost:8080

Log in with the DASHBOARD_USER / DASHBOARD_PASSWORD from your .env.

💡 Tip: set it up from a computer first. The in-app Settings page (feeds, categories, counts, theme) is much easier to configure with a full-size keyboard and screen. Get your feeds and look dialed in on a laptop or desktop, then just open the same URL on your Tesla or phone to read — all settings live on the server, so every device sees the same configuration.

How it works

              ┌──────────── one container (Next.js) ────────────┐
   RSS feeds ─┼▶ curator ──(writes)──▶ /data ──(reads)──▶ reader ┼─▶ you
 (per category)  fetch · filter ·      feed.json          in-place
                 dedup · rank ·        reactions.json     article view
                 summarize             config.json
              └─────────────────────────────────────────────────┘

• curator — runs inside the Next.js server (started from instrumentation.ts): fetches each category's feeds, drops paywalled/duplicate/old items, ranks (recency + your reactions + optional keyword boost), and writes feed.json on a cron schedule.
• reader — the Next.js app serving the feed, the in-place article reader, and the reaction/config APIs. Guarded by a single login.
• State persists to one Docker volume (/data): feed.json, reactions.json, config.json. (Earlier versions split this into separate worker/web containers; it's now a single process — see CURATOR_ENABLED if you run multiple replicas.)

Configuration

All via environment variables (see .env.example):

Variable	Default	What it does
DASHBOARD_USER / DASHBOARD_PASSWORD	admin / —	Login credentials
AUTH_SECRET	auto-generated	Session-cookie signing key (persisted to /data if unset)
WEB_PORT	8080	Host port for the reader
CRON	0 8,18 * * *	Curation schedule (cron)
TZ / TIMEZONE	America/New_York	Timezone for the schedule
RUN_ON_START	true	Curate once on startup (fresh content on first boot)
CURATOR_ENABLED	true	Set false to disable curation on this instance (multi-replica setups)
SEARXNG_URL	—	Optional self-hosted SearXNG base URL to augment RSS with web-search news (needs JSON format enabled)
LLM_PROVIDER	none	none | anthropic | openai for summaries

Everything else — feeds, categories, per-category counts, recency/retention, keyword boosting, accent color, and reading size — is configured in the in-app Settings page (⚙️ in the header) and stored in config.json on the data volume. No rebuild needed.

Customizing feeds

Open Settings (easiest from a computer — see the tip above) → add/remove/reorder categories and feeds, set per-category counts, pick a boost category + keywords (e.g. your town for "Local"), and choose the accent color and reading size. Changes to feeds apply on the next curation run; theme changes apply immediately. You can also import/export your feed list as OPML, or edit config.json directly.

Putting it behind a domain name (recommended)

Out of the box WattsNew listens on http://<host>:8080. For day-to-day use — especially from a Tesla or phone — it's much nicer behind your own domain with HTTPS (e.g. https://news.example.com). Put a reverse proxy in front of it.

WattsNew is reverse-proxy-aware: when it sees an HTTPS request (via the X-Forwarded-Proto header), the session cookie is automatically marked Secure. So just forward that header and you get hardened cookies for free.

nginx example (after pointing your domain's DNS at the host and issuing a cert, e.g. with certbot):

server {
    listen 443 ssl;
    server_name news.example.com;

    ssl_certificate     /etc/letsencrypt/live/news.example.com/fullchain.pem;
    ssl_certificate_key /etc/letsencrypt/live/news.example.com/privkey.pem;

    location / {
        proxy_pass http://127.0.0.1:8080;   # or the container's LAN IP
        proxy_set_header Host              $host;
        proxy_set_header X-Real-IP         $remote_addr;
        proxy_set_header X-Forwarded-For   $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;   # enables Secure cookies
    }
}

# Redirect plain HTTP → HTTPS
server {
    listen 80;
    server_name news.example.com;
    return 301 https://$host$request_uri;
}

Run one server block per domain if you host multiple readers. Prefer something turnkey? Caddy does the same with automatic HTTPS in two lines:

news.example.com {
    reverse_proxy 127.0.0.1:8080
}

(Caddy sets X-Forwarded-Proto automatically, so Secure cookies just work.)

Roadmap

• Config-driven feeds/categories via config.json + an in-app Settings UI
• Theme settings (accent color, font-size presets) in the UI
• OPML import/export
• Mobile/responsive layout (master-detail on phones)
• Prebuilt images on GHCR + GitHub Actions CI
• Multiple users — per-user logins & reactions; admin-managed accounts
• Read/unread, search, save-for-later
• PWA install

Deploying to a remote Docker host

Point your Docker CLI at the host (e.g. an SSH context) and run compose there:

docker context create myhost --docker "host=ssh://user@your-host"
docker --context myhost compose up -d --build

License

MIT.