Skip to main content

Stats dashboard (wolt stats)

wolt stats brings your Wolt order history home. One command downloads the wolt-stats dashboard bundle, syncs every order Wolt has on file into a local SQLite database, serves the dashboard at http://127.0.0.1:5173, and opens your browser. Nothing leaves your machine — the dashboard is a static SvelteKit app reading the SQLite file in the browser via sql.js.

wolt stats dashboard overview

Quick start​

# First run: download bundle, sync history, open dashboard
wolt stats

# Subsequent runs: incremental — only fetches new orders
wolt stats

# Re-open the dashboard without re-syncing
wolt stats --no-sync

# Force a full re-scan of every order
wolt stats --resync

# Serve without opening a browser (e.g. on a server)
wolt stats --no-open

What it tracks​

The sync persists everything Wolt returns from the order-history endpoints:

  • Catalog summaries — purchase id, payment timestamp, venue, currency, totals, short item summary, status, delivery method, main image.
  • Full order details — line items with options, payments (provider, card name, credits applied), delivery distance and base price, service fee, discounts, gift cards, order adjustment rows, creation/delivery times.

From those rows the dashboard derives spend by month, top venues, item leaderboards, weekday/hour-of-day patterns, average order size, delivery-fee breakdowns, and more. The full schema lives in internal/service/statssync/store.go.

per-venue and per-item drill-down

Where data lives​

~/.wolt/stats/
├── bundle/                       # downloaded dashboard release (versioned)
│   └── v0.2.1/                   # extracted SvelteKit static site
└── db/
    └── wolt-history.sqlite       # your sync target — read by the dashboard

The directory location is overrideable with --stats-dir or $WOLT_STATS_DIR.

Sync model​

The catalog phase pages backwards through /v1/order_history until it hits an order that's already in the local DB (incremental mode), then stops. The detail phase fetches /v1/order_history/purchase/<id> for every catalog row that doesn't yet have a orders table entry.

Two phases share a single inter-call pacer:

  • Baseline pacing is 1.1 s/call. Empirically this is at or under Wolt's per-token throttle ceiling.
  • Each HTTP 429 response bumps the pacer by +500 ms for the remainder of the run (capped at +5 s extra), so the sync settles into whatever sustained rate your account actually permits.
  • Individual 429 retries use exponential backoff (2s → 4s → 8s → …, capped 60s) and honor any Retry-After header.
  • If retries are exhausted on a single order, the run surfaces a resumable error pointing at where it stopped. Re-running picks up where it left off.

Flags​

FlagDefaultEffect
--no-syncfalseSkip the order-history sync, serve whatever is already in the local DB. Useful for re-opening the dashboard quickly.
--no-openfalseDon't launch the browser. Pair with --port for headless/server use.
--resyncfalseForce a full backfill instead of incremental. Use this if the catalog stop heuristics miss an order or if you've deleted/edited rows.
--no-check-updatesfalseSkip the GitHub release check. The cached bundle is reused as-is.
--bundle-versionlatestPin to a specific wolt-stats release tag (for example v0.2.0).
--port5173Preferred HTTP server port. Auto-picks +1 if the port is busy.
--stats-dir~/.wolt/statsOverride the install/data directory. Also reads $WOLT_STATS_DIR.

All global flags (--format, --verbose, --locale, etc.) apply too.

Privacy​

Your order history never leaves your machine. The dashboard is plain static files served from 127.0.0.1 — no telemetry, no third-party requests, no analytics. The only network traffic is:

  1. wolt-cli → Wolt's consumer API, with your saved access token, to fetch order history (same endpoints your browser already calls).
  2. wolt-cli → GitHub, once per run, to check for a newer dashboard release (suppressible with --no-check-updates).

The on-disk SQLite is owner-readable (0o600) under ~/.wolt/stats/db/.

Implementation notes​