Upload PDF bank statements. AI extracts and categorizes every transaction. Detect subscriptions, visualize spending flows, and get actionable insights — all running locally on your machine.
Features
Drag-and-drop bank or credit card statement PDFs. Claude extracts every transaction with date, description, amount, and category. Duplicate detection across overlapping statements.
AI classifies transactions into 71 categories across 16 groups. Searchable combobox for manual edits. Re-classify everything when the taxonomy improves.
Automatically detects recurring charges by normalized merchant name. See frequency, monthly cost, and yearly totals. Merge duplicates, dismiss false positives.
Bar charts, pie charts, and trend lines for spending analysis. A custom Sankey diagram visualizes money flow from income sources through category groups to individual categories.
Claude Haiku analyzes 6 months of data and generates observations about cross-category patterns, unusual spending, and actionable suggestions. Statistical trend detection too.
Everything runs on your machine. SQLite database stored locally. PDF statements never leave your computer. The only external call is to Claude for AI extraction.
Transaction Management
Filter by type, category, date range, or document. Inline category editing with a searchable combobox. Bulk select and delete. Pagination for large datasets. Search across all descriptions.
Reports
Summary cards for total spent, income, and averages. Spending over time bar charts, category breakdown pie charts, and spending trend lines comparing debits vs credits.
Money Flow
A custom Sankey diagram traces income from sources through category groups to individual categories. Net savings are shown as a separate node so you can see the full picture at a glance.
Insights
Claude analyzes months of spending data and surfaces patterns you might miss. Category trends, lifestyle inflation, recurring charge growth, and spending shifts — ranked by severity and recency.
Subscriptions
Automatically detects subscriptions and recurring payments by normalized merchant name. See frequency, average cost, monthly and yearly totals. Merge duplicates, dismiss false positives.
Spending trends and top transactions
Getting Started
$ git clone https://github.com/aromasca/expenlytics.git
$ cd expenlytics && npm install
$ echo 'ANTHROPIC_API_KEY=sk-ant-...' > .env.local
Used for PDF extraction, categorization, and insights. Get a key at console.anthropic.com.
$ npm run dev
Open localhost:3000 and upload your first PDF statement.
Built With
Next.js 16
App Router
React 19
Server & Client
TypeScript
Strict mode
Tailwind v4
Utility CSS
SQLite
better-sqlite3
Claude AI
Anthropic SDK
Recharts
Charts
d3-sankey
Flow diagrams
Open source, local-first, AI-powered. Clone the repo and upload your first statement.
View on GitHub