ReceiptRoller 1.4.193 Release Notes

In this 1.4.193 release, building on the theme we've been developing since the previous 1.4.168 release — "not just making data visible, but making decisions together with AI" — we've shipped a wide range of updates: a new MCP server for AI agent integration, a monthly report Intelligence Panel, ABC analysis by product, visitor analysis, product demand forecasting, churn risk scoring, year-over-year comparison with multi-store comparison, and multi-organization support for the native store app API.

You can use the service here: https://receiptroller.io

What's new

New MCP server for AI agent integration

We've launched a new MCP (Model Context Protocol) server that lets AI agents such as Claude query your ReceiptRoller store data directly. Ask natural-language questions like "How were sales last month?" or "Which customers look like they might churn?" and the AI can answer by referencing your store data as-is.

• The endpoint is https://mcp.receiptroller.io. Connect from any MCP-compatible AI client (Claude Desktop, Claude.ai, various IDE integrations, etc.)
• We've published a wide range of read tools for SNS integration (accounts, posts, analytics, comments) and sales analysis (KPIs, trends, advice, forecasts, products, visitors, churn, multi-store comparison)
• We've adopted a "user-scoped token" approach that lets a single access token cross-reference all organizations you belong to. You specify the organization when calling a tool
• Supports OAuth 2.0 + PKCE + Dynamic Client Registration, so setup on the AI client side completes with one click

Added an "Intelligence Panel" to the monthly dashboard

We've added an "Intelligence Panel" to the monthly finance/sales dashboard that automatically analyzes your store data and displays actionable insight cards. It updates automatically every morning, and you can also refresh it manually anytime via the "Regenerate" button.

• Trend analysis — Compares this month's sales against last month and the same month last year, with the current month's figure projected through month-end
• Month-end forecast — Forecasts month-end sales based on past results and day-of-week seasonality, with a confidence range shown as a band
• Anomaly detection — Automatically highlights days where sales deviated significantly from the usual trend
• Category mix shifts — Detects changes in the sales composition ratio of your main categories
• Product demand forecast — Shows projected quantities for the next 14 days for your top 5 best-selling products
• Churn risk — Scores "customers who look likely to churn" based on visit interval, visit frequency, and purchase amount
• Cards are automatically sorted by priority (needs action / caution / opportunity / informational)

Added ABC analysis by product and sales analysis for custom date ranges

We've added a screen that automatically classifies per-product sales into A/B/C ranks based on the Pareto principle. You can see at a glance which are your core products and which ones deserve more attention.

• Automatically classifies each product's sales amount, quantity, customer count, and gross profit into three ranks: A (cumulative 70%), B (cumulative 70-90%), and C (the remainder)
• Alongside presets like "this month," "last month," and "last 7/30/90 days," you can now specify a custom date range for analysis

New visitor analysis (time of day, day of week, new vs. repeat)

We've added a visitor analysis screen that visualizes visit patterns from your receipt data. See busy time slots and new-vs-repeat trends at a glance.

• Displays a visit heatmap by time of day and day of week
• Shows bar charts of visit count, average spend per customer, and sales by time slot, plus the same metrics by day of week
• Shows monthly trends by new vs. repeat customers as a stacked graph, so you can track changes in your repeat rate
• Also displays a reference conversion metric relative to visitor count

Added churn risk scoring and product recommendations

We've added a feature that automatically scores "customers who look likely to churn" based on visit history. Since it's linked with CRM customer data, it ranks customers by risk based on visit interval, purchase amount, and last visit date.

• Calculates churn risk from three metrics: last visit date, visit frequency, and purchase amount
• Based on past purchase history, suggests products this customer is likely to buy next from patterns in commonly-bought-together items
• Customers above a certain churn risk threshold are also shown as "churn risk cards" in the Intelligence Panel

Added year-over-year comparison, custom date ranges, and multi-store comparison

Both organization reports and store reports now support comparison against the same month last year, custom date range selection, and overlaid multi-store comparison.

• The monthly trend graph now overlays last year's values on the same month, and you can check the year-over-year change rate via tooltip
• In organization reports, you can select multiple stores from the store picker to overlay and compare
• The same comparisons are also available via the API / MCP tools

New sales-related REST API (/api/v1/sales/*)

We've published a REST API that lets you fetch the sales analysis data shown on the store management screen directly from an external system or mobile app. The scope is sales.read.

• Published endpoints: /api/v1/sales/kpis, /trend, /advice, /forecast, /products, /visitors, /churn, /recommendations, and /multi-trend
• The same data is also available from the MCP server's sales_get_kpis / sales_get_trend / sales_get_advice and related tools

Multi-organization support for the native store app API

For owners who operate multiple stores or multiple organizations, we've extended the API so a single access token can cross-reference all of your organizations. This lays the groundwork for native store app development on Android / iOS.

• Key APIs for stores, POS, sales, SNS, customers, and chat can now switch the target organization via the ?organizationId= query parameter
• GET /api/v1/me/organizations returns the full list of organizations the user belongs to
• Every call goes through membership validation for the target organization, so there's no risk of seeing another organization's data

Historical data backfill and month-end projection fix

To improve the accuracy of store data, we've added a backfill feature that reconstructs past monthly aggregates from POS transaction data. Stores that newly connect their POS can now view past trends right away.

• The current month in the monthly trend graph is now compared against last month after being automatically projected through month-end. This eliminates the false impression mid-month that sales "dropped from last month"
• Fixed an issue where the monthly trend graph's bars didn't scale correctly relative to their values

Other fixes and improvements

• Fixed the navigation link to a store's "Reports" page
• Sped up help page rendering
• Fixed an issue where some overnight auto-updated advice failed to display

Related help & resources

• Top
• For stores
• For developers
• MCP server endpoint
• Feature requests (send us your feedback)
• Previous release notes (1.4.168)
• Android app
• iOS app

What's next

In upcoming releases, we're planning to work on: the full release of native store manager apps for Android / iOS, additional write tools for the MCP server (post scheduling, customer contact, etc.), automatic campaign suggestions based on visitor analysis, real-time notification integration when receipts are issued, and the full release of the store monthly dashboard. If there's a feature you'd like to see, please let us know anytime via the feature request page.

Thank you for using ReceiptRoller. If you have any comments or requests, please reach out via "Contact Us" in the store management screen.