Inspiration
We’ve all been there, you check your bank account on Sunday and realize your "little treats" added up to a big problem. Most budget apps fail because they require you to open a new app and manually log data. We wanted to build something that meets you where you already are. Lanny is a zero-friction AI assistant that lives directly in your iMessages, turning your spending into a fun but high-stakes challenge.
What it does
Lanny is a proactive financial guardian. It tracks every transaction in real-time, answers complex questions about your habits "How much did I spend on oat milk lattes this month?", and hosts the "Stakes Challenge." If you set a budget and break it, Lanny uses its Stripe and Every.org integrations to automatically donate $10 of your money to charity. It turns bad habits into good deeds.
How we built it
- KnotAPI : Our engine for user's transaction data, allowing Lanny to see spending as it happens.
- K2 Think V2 : An open-source LLM that analyzes transaction history to provide personalized, witty feedback and saving strategies.
- Photon Spectrum : Used to integrate the AI directly into iMessage, removing the "app friction" that usually kills budgeting habits.
- Stripe : Handles the "accountability fee" by processing automated donations when users fail their challenges.
- Every.org : The held Stripe payment is captured and automatically donated to a charity (default: Feeding America) via the Every.org API. The user's money goes to a good cause instead of just being a penalty!
Challenges we ran into
As first time KnotAPI users, mapping real-time bank data to a conversational AI model was quite a challenge. We attended workshops and iterating on prompt engineering to ensure Lanny could distinguish between a necessary bill and an impulse buy.
Accomplishments that we're proud of
We successfully built a full-circle accountability loop. Lanny isn't just a "tracker", it's an active participant in your financial life. We’re particularly proud of our "invisible UI," where the entire experience happens through a familiar chat bubble.
What we learned
We gained deep experience with KnotAPI’s webhooks and discovered how to use Photon to bridge the gap between high-level AI models and daily communication tools.
What's next for Lanny
We’re going beyond tracking. The next version of Lanny will proactively scrape the web to find coupons and better deals for the stores you already frequent, saving you money before you even hit "purchase."
Built With
- claude
- cursor
- javascript
- k2think
- knot
- photon
- postgresql
- typescript
Log in or sign up for Devpost to join the conversation.