RReceipt OCR App
Open App

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

[Unreleased]

Added

  • Multi-receipt batch upload with queue UI — select multiple files at once, sequential processing with per-file progress indicators (uploading, OCR, classifying, done/error), overall X/Y progress counter
  • Failed files show per-file errors without blocking other uploads

Changed

  • Migrated from HTTP API layer to direct D1 adapter (@marlinjai/data-table-adapter-d1) for database access
  • Classification failures are soft — row saved with Pending status, no upload blocking

[0.5.0] - 2026-02-28

Added

  • Liquid Glass UI aesthetic with aurora background, glass-panel surfaces, and backdrop-filter blur
  • Direct-to-DB receipt uploads — receipts persist immediately on upload
  • Row selection with backspace/delete keyboard deletion
  • Column alignment and keyboard navigation (arrow keys, Tab)
  • Improved receipt name extraction with 3 item detection patterns (price-based, quantity-based, SKU-based)
  • Broader German/English noise filtering and deduplication for receipt parsing

Changed

  • Bumped data-table packages to ^0.2.0 for liquid glass UI and keyboard nav support

Fixed

  • Refocus table on background click to restore keyboard navigation
  • Single-click cell editing and show ungrouped rows (data-table-react 0.1.3)

Removed

  • Receipt-store polling module — replaced by immediate database persistence

[0.4.0] - 2026-02-27

Added

  • Database integration as persistent storage backend for all environments
  • SKR03 accounting categories (Bewirtung, Reisekosten, Bürobedarf, etc.) for German Vorkontierung
  • Konto column with SKR03 account numbers
  • Notion-style grouping by category/konto in table views
  • German vendor and keyword inference maps for category detection

Fixed

  • Exclude clearify.config.ts from TypeScript build check

[0.3.0] - 2026-02-20

Added

  • Server-side OCR API route using Google Cloud Vision
  • Two-phase upload flow — Storage Brain upload then OCR via Vision API
  • D1 adapter support with memory fallback for local dev
  • OpenNext Cloudflare deployment configuration with D1 binding
  • Storage Brain SDK workspace support (v0.5.0) with automatic workspace scoping
  • Marketing landing page with hero, features, and how-it-works sections
  • Custom domain route for receipts.lumitra.co
  • Deployment to Cloudflare Workers

Changed

  • Moved upload page from / to /app, dashboard from /dashboard to /app/dashboard
  • Replaced file: links with published npm packages (data-table v0.1.x)
  • Decoupled OCR types from Storage Brain SDK into app-owned types

Fixed

  • Turbopack CSS import panic for file-linked packages
  • Turbopack dev compatibility with dist/ CSS import path
  • Updated storage-brain-sdk to v0.4.0, removed deprecated OcrResult type

[0.2.0] - 2026-02-19

Added

  • Dashboard page with Data Table integration (Table, Board, Calendar views)
  • Receipts table initialization with schema and views
  • Upload-to-dashboard flow with receipt data ingestion
  • Intelligent receipt field extraction pipeline (vendor, amount, date, category)
  • Regex/heuristic parsing for US and European number/date formats
  • Spatial vendor detection from OCR blocks
  • Keyword-based category inference against ~50 known vendors

[0.1.0] - 2026-01-11

Added

  • Initial Next.js application setup
  • Drag & drop receipt upload component with progress tracking
  • OCR result display with confidence score and copy functionality
  • File preview for images and placeholder for PDFs
  • Storage Brain SDK integration for file uploads with invoice OCR context
  • File details panel with status indicators