Wispr Flow: Dictation That Works EverywhereStop typing. Start speaking. 4x faster.
Promoted
Maker
๐
Hey Product Hunt! ๐ I'm Tejas, maker of Beancounter.
Every expense tracker I've ever used died the exact same death: I stopped logging.
Not because I didn't care about my money, but because logging a $4 coffee usually takes six taps across three screens. Doing that at a counter while someone waits behind you is genuinely not worth $4 of effort, so you skip one, then a week, and suddenly the app is an abandoned graveyard.
The problem was never motivation, it was friction. So I built Beancounter, a Telegram bot where the interface is the one you already open 50 times a day.
How it works:
- โก Zero-Friction Logging: You type purse 5.2 or 20 credit for groceries. The bot logs it and reacts with a quiet ๐ on your message. No clutter, no essay-long replies.
- ๐ง Hybrid Parsing: 95% of standard messages are parsed instantly by plain Python with zero latency. For messy/ambiguous text (5 in purse and 20 from credit), it quietly hands off to an LLM fallback.
- ๐ท๏ธ Smart Nicknames: You don't have to remember exact wallet names. A wallet can have multiple nicknames (cash, money, purse), so you just write however you naturally think.
- ๐ Built-in Mini App: Run /app to open a clean ledger directly in Telegram, complete with spend-by-weekday stats, hourly heatmaps, and trend lines.
- ๐ต๏ธโโ๏ธ Honest Budgeting ("Untracked Drift"): When you correct a wallet balance with /setwalletamnt, the gap is tracked as "untracked drift"โthe money that moved without being logged.
- ๐ Private & Open Source: 100% open source, zero bank-linking, no ads, and fully self-hostable with a local SQLite database you own.
I built this primarily to fix my own bad tracking habits. I'd love to hear your feedback, feature ideas, or brutal UX critiques! Drop a comment below and let me know what you think. ๐งฎ๐ธ