FunPlaneViewer

I run an ADS-B feeder using adsb.im with SkyStats enabled. SkyStats uses the plane-alert-db to log interesting aircraft spotted by your feeder, but its built-in GUI is designed around showing just a handful of recent planes. It can technically display more, but it lacks the filtering and layout to make that practical.

This project is a vibecoded fix meant for my personal use. It uses the same SkyStats backend as the official GUI, but presents the full history in a much more readable way.

It's also optimised for mobile since that was my main priority.

Browsing

• Military, government, and civilian tabs backed by the SkyStats interesting-aircraft endpoints.
• SkyStats' fourth feed, police, is folded into the government tab rather than getting a tab of its own — law-enforcement airframes are state-operated, not armed forces. They keep their Police category, so the category chips still let you filter them out on their own. Backends too old to serve /api/stats/interesting/police are handled gracefully: the feed is skipped and everything else loads as normal.
• Grid and table views for stepping through a large aircraft history.
• Search across type, operator, callsign, ICAO, registration, category, and tags.
• Sort by last seen, type, operator, or category.
• Category filter chips with live counts.
• Aircraft detail modal with image gallery, tags, metadata, and last-seen time.
• A "Track" button on each card opens that airframe on globe.airplanes.live by hex, in a new tab.
• Image fallback cards for aircraft without photos.
• 24-hour time formatting.
• Mobile-friendly layout with a collapsing header.
• Warning banner when the data on screen is known to be incomplete: if a live section starts returning far fewer aircraft than it used to, or the shared backup on the Pi can't be loaded. A short list otherwise looks identical to a quiet week, which makes an upstream problem read as deleted aircraft.

Stats dashboard

• A Stats tab with a switch between two views: "Backend stats" (feeder-wide) and "My aircraft" (your logged interesting aircraft).
• Backend stats come from the SkyStats /api/stats endpoints:
  • Totals for unique aircraft seen and flights tracked.
  • Top aircraft types, with toggles for metric (flights flown or unique aircraft) and timeframe (24h, month, year, all).
  • Top airlines, busiest routes, top domestic and international airports, and origin and destination countries.
• My aircraft stats are computed in the browser from the logged mil/gov/civ lists:
  • Activity summary: aircraft logged, new in the last 24h, busiest hour and day, and the last-seen date range.
  • Breakdown by group (military, government, civilian — police counted under government).
  • Seen by hour of day and by day of week (based on each aircraft's last-seen time).
  • Top operators, types, tags, and categories.
• Dependency-free CSS bar charts that pack neatly and stay readable on mobile.

Import and export

• An Export dialog offering three scopes, each with a live count of what it will write, and an "include tags" toggle:
  • This section — what the open tab is showing, search and category chips included.
  • All sections — Military, Government and Civilian in one file, filters ignored.
  • One category — a single category on its own (Oxcart, RAF, As Seen on TV...), gathered across all three sections, with a filter box for finding it.
• CSV and JSON import for restoring backups. Each row is routed by its own group column, so a combined export lands back in the sections it came from; files without a usable group fall back to the section named in the filename.
• Imported backups are merged into the live receiver feed instead of replacing it.
• Imported backup data persists locally in the browser, with optional shared persistence on the Pi (see server/README.md).
• Server-side backups are versioned before every overwrite, and a save that would wipe a section that still has aircraft is refused (see Backup safety).
• Optional import enrichment from plane-alert-db:
  • pulls missing image links from plane_images.csv
  • pulls missing tags and metadata from plane-alert-db.csv
• In-app import prompts instead of browser popups.

Daily backups

• The Pi sidecar snapshots every logged aircraft once a day at a fixed hour and keeps the last 30 days.
• Runs on the server, so it happens whether or not anyone has the GUI open. A Pi that was off at the scheduled hour takes its backup when it boots.
• The Daily backups tab in the hidden menu lists what's stored and restores any of them in one click, plus a Back up now button. Each row carries a bar showing the military / government / civilian split, so a snapshot that lost a whole section is obvious without opening it.
• Restoring merges the snapshot back over the live feed, so aircraft the feeder still reports stay current and anything missing comes back.
• Requires the Flask sidecar (see server/README.md).

Settings (triple-click the title)

App-wide preferences only, as three tabs:

• Backend URL — optional Tailscale-friendly override for accessing the GUI from outside the local network, with the address actually in effect and a live round-trip check below the field.
• Daily backups — browse and restore the automatic daily backups stored on the Pi.
• Update app — the installed version against whatever is on GitHub main, the recent commit subjects, and an "Update & reload" button that pulls the latest index.html so you don't need to ssh in to deploy a change (requires the Flask sidecar, see server/README.md). The version check is best-effort: without a route to GitHub the tab still offers a blind update.

Photo links (triple-click an aircraft photo)

Scoped to one aircraft, so there's no search step to get through first:

• Add manual per-aircraft image links, saved either locally or on the Pi for everyone.
• One-click JetPhotos lookup by registration, plus Google and Planespotters lookups by ICAO hex.
• Change aircraft opens the picker if you want to edit a different one without leaving the dialog.

The SkyStats backend defaults to returning only 5 planes, so you need to raise that limit before this GUI is useful. Pick one of the two options below.

1. Install dependencies: pip install requests colorama
2. Run patch.py
3. Enter your feeder URL when prompted (default adsb-feeder.local:5173)
4. Enter the limit you want to set (default 9999)
5. If the request succeeds, you're done.

1. Open SkyStats and go to Settings.

2. Open your browser's network tab (F12 then Network).

3. Change "Interesting Aircraft - Number of rows to display" to any value (the exact number doesn't matter).

4. Find the request that was triggered, right-click it, and copy as cURL.

5. Paste it into an API client like Postman.

6. In the request body, set interesting_table_limit to a large number like 99999. The full body should look like:

   {"route_table_limit":"5","interesting_table_limit":"99999","record_holder_table_limit":"5","disable_planealertdb_tags":"false"}

7. Send the request. If it succeeds, you're done.

By default the GUI talks to the SkyStats backend at the same hostname the page was loaded from, on port 5173. If you open the GUI via your feeder's Tailscale hostname or tailnet IP, the API calls automatically go over Tailscale too, as long as port 5173 is reachable on that host.

If your setup needs a different backend address (for example, the GUI is hosted on a different machine than the feeder, or you're proxying SkyStats through Tailscale Serve), open the hidden menu (triple-click the title), stay on the Backend URL tab, and enter the full URL, for example:

• http://my-feeder.tailnet-name.ts.net:5173
• https://feeder.tailnet-name.ts.net

The value is stored in your browser's local storage and used for all subsequent requests. Click Reset to default to go back to the automatic behavior. Both buttons leave the dialog open so the reachability line underneath can tell you whether the new address actually answers.

If the GUI loads but the aircraft list doesn't, the error message includes a Change backend URL button that opens the same tab.

Note: if you access the GUI over HTTPS (for example via Tailscale Serve with TLS), the backend URL also needs to be HTTPS. Browsers block mixed HTTP/HTTPS requests.

A small Flask sidecar adds shared server-side persistence for manual image links and imported backups, the automatic daily backups described above, plus a self-update button in the hidden menu that pulls the latest index.html from GitHub. It's strictly optional; without it the GUI works exactly as before, minus the daily backups.

See server/README.md for setup.