A growing set of tools for game developers:
- Release Calendar — every upcoming Steam game plotted by release date, so you can find a launch day with fewer games and fewer high-wishlist competitors.
- Tag Explorer — competition broken down by Steam tag/genre: which genres are most crowded with upcoming releases, when the crowd lands (monthly histogram), and the biggest competitors in each genre.
Live: https://egormagurin.github.io/SteamUtil/ (static build, auto-updated every 3 hours)
npm install
npm startThen open http://localhost:3000. On first launch the server fetches the
upcoming-games list from Steam in the background (a progress bar shows in the
UI) and caches it to data/upcoming.json. Subsequent starts load instantly from
that cache.
- Source. Steam's store search endpoint with
filter=popularcomingsoon(Games category only). This list is ordered by Steam's popularity ranking, which is driven by wishlist activity. - The wishlist proxy. Steam does not publish exact wishlist counts
anywhere public. Two free proxies are used instead:
- Popularity rank — a game's rank in "Popular Upcoming" (rank #1 is the most-wishlisted upcoming game). Captured for every game.
- Follower counts — a game's community-group member count
(
steamcommunity.com/games/<appid>/memberslistxml) is public and tracks wishlists closely. Shown as★ followersplus a rough wishlist estimate (followers × multiplier, multiplier adjustable in the UI, default ×10). The follower→wishlist ratio varies a lot per game, so the estimate is shown as an "≈" guide, never a hard number. The Biggest competitors panel ranks upcoming games in your window by followers.- Coverage note: Steam hard-caps this endpoint at ~60 requests per IP per window, so enrichment is limited to the top ~60 games by popularity — which are exactly the high-wishlist titles that matter for competition. Beyond that, requests just stall and get dropped.
- Competition heat. Each calendar day is colored by an index that weighs
high-wishlist games (top-N rank, configurable) more heavily than small ones:
index = totalGames + highWishlistGames × 4. Cooler = better. - Caching. Results are cached for 12h. Use ↻ Refresh data (or pick a larger fetch depth) to pull more games or get fresh numbers.
- Rate limits & resume. Steam rate-limits bursty requests (HTTP 429). The
fetcher backs off and retries automatically, and persists progress after every
page to
data/upcoming.partial.json. If a fetch is interrupted (rate limit, crash, restart), the next refresh resumes from where it stopped instead of starting over — partial progress is never wasted. Partials older than 1h are discarded (the popularity order may have shifted). The last complete dataset stays indata/upcoming.jsonand keeps serving the UI until a refresh finishes.
- Exact-date games appear on the calendar. Steam dates like
Jul 9, 2026. - Coarse dates (
Q3 2026,Jul 2026,2026) can't be placed on a specific day, so they're summarized in the sidebar. These still compete. - A massive spike at month and quarter boundaries (e.g. Jun 30, Sep 30) is normal — many developers default to end-of-period dates. Mid-month and mid-week days are usually far quieter.
- Fetch depth controls how many of the most-wishlisted upcoming games are pulled (2k / 5k / 10k). Higher depth = more complete long-tail counts for dates further out, at the cost of a slower refresh. Steam currently lists ~48k upcoming games total.
server.js Express server: API + static hosting + background fetch job
src/steam.js Steam fetch/parse/cache + release-date normalization
public/ Frontend (no build step): index.html, app.js, style.css
scripts/ build-data.mjs (static data for Pages), serve-static.mjs
.github/workflows/ deploy.yml — builds data + deploys to GitHub Pages
data/ Generated backend cache (gitignored)
GitHub Pages serves static files only — it can't run the Node backend. So the
published site uses a static snapshot: a GitHub Actions job fetches the
Steam data, writes public/data/upcoming.json, and deploys public/ to Pages.
The frontend auto-detects there's no backend and loads that file instead of the
API (the Refresh button is disabled and shows "Auto-updated every 3h").
-
Workflow: .github/workflows/deploy.yml runs on every push to
main, every 3 hours (xx:17 UTC) to refresh the data, and on manual dispatch. It auto-enables Pages on first run. -
Data size: controlled by the
DEPTHenv in the workflow (default 3,000). -
Follower enrichment: tuned by
FOLLOWERS_TOP(default 60, the Steam per-IP cap),FOLLOWERS_CONCURRENCY(1), andFOLLOWERS_THROTTLE(ms). Each game is one extra request; staying at/under ~60 keeps coverage high. -
Preview the static build locally:
npm run build:data # writes public/data/upcoming.json (set DEPTH to taste) npm run serve:static # serves public/ with no API at http://localhost:4000
(Local
npm startstill runs the full live backend on :3000.)
| Method | Route | Purpose |
|---|---|---|
| GET | /api/data |
Current cached dataset |
| POST | /api/refresh?depth=N&followers=M |
Start a background fetch (games + followers) |
| GET | /api/status |
Progress (incl. phase) + cache freshness |
- Wishlist Trends: per-weekday and per-month historical release-volume trends.
- Estimated wishlists shown directly on calendar day cells.
- Combine the calendar + tag filter (low-competition days within your genre).