Scan a bank statement in CSV, PDF, or image form to find zombie subscriptions and price hikes.
Generate a ready to send cancellation email for a confirmed unwanted charge.
Schedule a monthly automated scan that emails a report without needing to confirm anything by hand.
Run LeakLens through an MCP server so an AI assistant can scan statements and ask you to confirm findings in chat.
| klncgty/leaklens | 0xallam/my-recipe | 0xhassaan/nn-from-scratch | |
|---|---|---|---|
| Stars | 0 | — | 0 |
| Language | Python | Python | Python |
| Last pushed | — | 2022-11-22 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 2/5 | 2/5 | 4/5 |
| Audience | general | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires your own OpenRouter API key for merchant name normalization, since detection only sends merchant names, not full statements.
LeakLens is a command line tool that reads your bank or credit card statements and looks for money that is quietly slipping away. It checks for zombie subscriptions you forgot to cancel, prices that crept up over time on a recurring charge, and duplicate charges billed twice on the same day. Each finding comes with a confidence score, an estimate of how much it costs you per year, and an option to generate a ready to send cancellation email. You can feed it statements as CSV files, PDFs, or even photos and screenshots in formats like JPG or PNG. If the text in a scanned document is hard to read directly, it falls back to a vision model to pull out the numbers. The detection itself runs locally using pandas, and only merchant names are sent out to a language model through your own OpenRouter API key, so your financial details stay on your machine. The tool also recognizes statement formats from several Turkish banks, including Is Bankasi, Garanti BBVA, Yapi Kredi, Akbank, Ziraat, and Enpara, and maps their column names automatically. When a finding is uncertain, LeakLens pauses and asks you to confirm it before moving on, a pause and resume flow built with LangGraph that works the same whether you are using the command line, an MCP server, or a future web interface. Every scan is saved locally in a SQLite database so you can compare new results against past runs and see what changed. Beyond manual use, LeakLens can run without any human input for automated setups, such as a monthly cron job that emails a report, or being called through an MCP server so an AI assistant like Claude Desktop, Cursor, or VS Code can scan a statement and ask you to confirm findings directly in chat. It is released under the MIT license.
A local first command line tool that scans your bank statements for forgotten subscriptions, sneaky price increases, and duplicate charges, and estimates how much money each one is costing you.
Mainly Python. The stack also includes Python, LangGraph, pandas.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly general.
This repo across BitVibe Labs
double-check against the repo, no cap.