Eskopia

Explain one alert (alert_id from get_alerts): for rule alerts, the rule's name and conditions and a summary of the triggering event or order (no customer identity); for anomaly alerts, the related detected anomaly and whether that condition is still active; plus delivery status per channel. Use for 'explain this alert', 'is it still happening', then investigate_store_change with the alert_id for 'what happened around then'.

get_alert_details

Alerts Eskopia sent for this store (rule matches on events or orders, anomaly alerts, digests), one entry per alert with delivery channels and status. Only anomaly alerts carry a severity; rule alerts are 'unclassified'. An admin sees all of the store's alerts; a regular user sees only alerts delivered to them. Use for 'what alerts do I have', 'did Eskopia alert me today', 'latest alert', 'critical alerts'. Use get_alert_details to explain one. This connection is read-only and cannot create or change alerts: if asked to, say so and use the result's alert_setup field, which lists exactly where and how alerts are set up in the Eskopia web app. Describe only what it lists; don't invent other fields.

get_alerts

Anomalies detected by Eskopia's deterministic store monitors (no_orders, revenue_drop, traffic_drop, checkout_completion_drop, tracking_gap, abandoned_checkout_spike, webhook_gap): status (active/resolved), severity, when detected, last confirmed, when resolved, evidence, and how many times the same detector fired before (last 30 days). Each detector is a fixed rule with minimum samples, not a statistical model; there are no confidence scores. Use for 'anything unusual', 'what anomalies today', 'is it still happening', 'has this happened before'. Records marked retroactive came from Eskopia re-checking its stored history. current_checks gives each detector's state right now: an empty anomaly list with checks in insufficient_data means Eskopia could not check, not that nothing is wrong.

get_anomalies

Group the store's visitors, activity, orders and revenue by one dimension, optionally narrowed by filters. Use for any 'by X' or 'from X' question: visitors or orders from a country, cities within a country (dimension=city with a country filter), traffic sources, mobile vs desktop, browser, landing pages, most viewed pages, new vs returning visitors or customers, and trends per hour, day, week or month. Each row has visitors, events, product viewers, add-to-cart visitors, checkouts started, orders, revenue, conversion rate, average order value and share of total. totals covers everything matching the filters, so a filter alone answers 'how many from X'. Countries are accepted as English names or ISO codes. Orders and revenue are null for page, product, collection and event_type (use get_product_sales for what each product sold). compare_to adds each row's change versus earlier windows.

get_breakdown

When the store is busiest: a weekday by hour-of-day grid (store timezone) for one metric over the last N days, with the busiest and quietest slots, totals per weekday and per hour. Metrics: visits (distinct visitors), orders, revenue, checkout_starts. Use for 'when are my peak hours', 'best day of the week for sales', 'when should I run a promotion or staff support', 'what time do people usually buy'.

get_busy_times

Checkouts started in a window and what became of each: summary (started, completed, abandoned, still in progress, completion rate, total abandoned cart value, where non-completed checkouts stopped, abandoned checkouts by country) and a list of individual checkouts with start time, status, furthest stage reached (started, contact info, shipping info, payment info, completed), cart value, country, city and minutes active. Filter the list by status; narrow everything by country or minimum cart value. Use for 'show me abandoned checkouts', 'how much money was left in abandoned carts today', 'biggest abandoned cart', 'is anyone checking out right now', 'at which step do people give up'. For step-by-step conversion rates and comparisons use get_funnel. Shopper names and contact details are never returned.

get_checkouts

Shopper funnel for a window: visitors -> viewed a product -> added to cart -> reached checkout -> completed purchase (distinct visitors), plus the checkout-stage funnel (started -> contact -> shipping -> payment -> completed) for checkouts started in the window, each step's conversion and drop-off, the biggest drop, and completed / not completed / still in progress checkouts. With compare_to it reports each step's change in percentage points and the largest worsening step. Use for 'where are customers dropping out', 'adding to cart but not buying', 'is the problem before or during checkout', funnel comparisons.

get_funnel

What's happening on the store right now: visitors active in the last 60 seconds by activity (browsing, viewing a product, cart, checkout), top countries and cities, what happened in the last N minutes (visitors, events by type, orders, revenue), up to 5 recent orders (no customer details), open checkouts, and whether this window is below/within/above the same window on recent same weekdays. Use for 'what's happening now', 'is anyone shopping', 'any orders recently', 'last 10 minutes / last hour', 'where are my current visitors from', 'is it unusually busy or quiet'.

get_live_activity

Orders received in a window: a summary over every matching order (count, revenue, average, median, largest and smallest order, items sold, first-time vs returning customers, top countries, how many came through CSV import or without a pixel-tracked checkout) and a list of individual orders with time, amount, currency, country, region, city, number of items, the products in the order (name, variant, quantity) and first-time/returning. Filter by country, city, amount range or customer type; sort newest, oldest, largest or smallest. Use for 'show me today's orders', 'biggest order this week', 'orders over 500', 'orders from Kuwait', 'how many first-time customers'. Customer names, emails and order numbers are never returned (they are in the Eskopia web app and Shopify admin). For totals per product use get_product_sales.

get_orders

Store performance for a time window, optionally compared with earlier windows: visitors, orders, revenue, conversion rate (orders / visitors), average order value, checkouts started and checkout completion rate, with percent changes. With compare_to=last_4_same_weekdays it also gives the min/average/max range and whether each metric is below, within or above that range (Eskopia's notion of 'normal for this time'; not a statistical baseline). Use for sales/orders/conversion/AOV questions, 'today vs yesterday', 'this week vs last week', 'is this normal for this time of day', executive summaries. For the same numbers split by country, city, traffic source, device or page, use get_breakdown. series=hourly (windows up to 72h) or daily adds a time series.

get_performance