Sarim Ahmad

PennyWise

PennyWise turns the transaction SMS that banks already send into categorised, searchable financial records. Parsing, storage and the optional AI assistant (a local Qwen 2.5 model via LiteRT-LM) run on the device; there are no accounts, ads, analytics or tracking SDKs. The only network use is optional: exchange-rate lookups for multi-currency features, and a user-initiated 'report a parsing problem' flow that sends the SMS text the user chooses to share. Every feature, including the Pro tier, is free in the F-Droid build. The parser engine (parser-core, pure Kotlin) supports 157 banks across 24 countries and ships in apps on Google Play, F-Droid and the iOS App Store. As of September 2026: about 1,300 installed users on Google Play (83% in India), 1,440 monthly active devices (up 35% over the previous 28 days) and a 4.4-star rating, with no paid marketing. On GitHub, since July 2025: 559 stars, 147 forks, 21 contributors and 8,000+ APK downloads from releases. Other projects build on it, e.g. Cashiro (AGPL-3.0, 188 stars), an expense tracker derived from PennyWise. Banks change their SMS formats often, and when they do, transactions go unrecorded until the parser is updated. Funding would pay for that ongoing maintenance: fixing and adding bank parsers, regression tests, releases, and reviewing community contributions.