AuctionScout

…

Server settings

Saved to the server's .env file. Changes apply after a restart. Settings can only be changed from the PC running the server.

Web server

Port

Access

HTTP→HTTPS redirect port

Schedule Nightly auto-refresh

Runs a full all-counties refresh once a day at this time (server's local clock). The server must be running — as a Windows service it always is.

Sale-day results check

The e-mailed report when a refresh ends (when it is sent, and to whom) is set in the customer portal: Settings › Refresh report.

Remote access (Cloudflare) Access team domain Application Audience (AUD) tag Allowed e-mails (optional) Who may come in through Cloudflare

Needed for https://app.<your domain> through the Cloudflare Tunnel and for admin controls in the customer portal: the app checks Cloudflare's sign-in with these. Restart after saving.

HTTPS certificate Certificate type

Counties

Unchecked counties are hidden and skipped by "All counties" refreshes. Their existing data is kept.

Data APIs RentCast API key (value estimates & comps) Google Maps API key (inline Street View photos) Census API key (free — ZIP median home values when a property has no estimate; get one) Auto address labels: max Google lookups per map view (buildings without an OpenStreetMap address are reverse-geocoded via Google; cached after. 0 = free OSM labels only)

Min comps required

Comp radius (miles)

Comp max age (days)

Valuation mode

Monthly API budget (hard stop)

RentCast billing period starts on day (1–28)

Only value sales within (days)

Auto-value top N per refresh (AI-ranked)

Rent estimates Property details: taxes, HOA fee and last sale

The value itself is 1 call; each switch on adds one.

AI analysis AI server URL (Ollama: http://localhost:11434 — or an OpenAI-compatible /v1 endpoint) Model name Vision model — reads the handwritten values on posted appraisal forms (Louisville). Ollama: ollama pull qwen2.5vl:7b; blank = forms are linked, not read

AI pre-screen before valuing

Min vet score to value (0–10)

Parallel AI streams (1–8) — batches vetted concurrently. For a real speedup, also set OLLAMA_NUM_PARALLEL on the Ollama server (same number) and restart it; otherwise extra streams just queue.

Scraper output✕