RecallPay

Inspiration

The frustration of losing receipts, forgetting about subscriptions, and missing warranty expiration dates inspired Receipt Recall. We wanted to create a smart financial memory that automatically captures and organizes every transaction, making it impossible to lose important purchase information again.

What it does

Receipt Recall automatically captures payment transactions through Flowglad integration, stores them in a searchable database, and uses Honcho AI's conversational memory to provide intelligent search capabilities. It tracks receipts, warranties, and subscriptions while generating smart reminders for warranty expirations, return windows, and subscription renewals. Users can search their transaction history using natural language queries like "What did I buy at Amazon last month?"

What makes it unique

Unlike traditional receipt apps that require manual scanning or photo uploads, Receipt Recall captures transactions automatically at the point of payment through Flowglad integration, zero manual work required. While competitors focus on static storage, we leverage Honcho AI's conversational memory to understand context and learn spending patterns over time. Our natural language search lets users ask questions like they would to a friend, not through rigid filters and categories. We're the only solution that combines automatic payment capture, AI-powered memory, and proactive warranty/subscription management in one seamless platform. This "set and forget" approach means users never lose another receipt, miss a warranty claim, or forget about a trial subscription again.

How we built it

We built a full-stack TypeScript application using React 18 and Express.js, with PostgreSQL for data persistence. We integrated Flowglad's payment API to automatically capture transaction data and Honcho AI for intelligent memory and search capabilities. The frontend uses Tailwind CSS and shadcn/ui for a clean interface, while the backend implements secure authentication with passport.js and session management.

Challenges we ran into

Switching from Replit Auth to a custom username/password authentication system required rebuilding the entire auth flow. We encountered session table conflicts between Drizzle ORM and connect-pg-simple that needed careful resolution. Integrating Honcho AI's JavaScript SDK required adapting their Python-focused documentation to work in our Node.js environment.

Accomplishments that we're proud of

Successfully integrated two external APIs (Flowglad and Honcho) to create a cohesive transaction memory system. Built a secure authentication system from scratch with proper password hashing and session management. Created an intelligent reminder system that automatically identifies expiring warranties and closing return windows.

What we learned

We learned how to properly implement secure authentication with passport.js and PostgreSQL session storage. We discovered the importance of proper error handling when integrating multiple external APIs. We gained experience adapting AI memory systems to work with financial transaction data for practical user benefits.

What's next for RecallPay

Implement OCR capabilities to extract text from uploaded receipt images. Add email integration to automatically capture emailed receipts from online purchases. Build mobile apps for on-the-go receipt scanning and transaction tracking. Expand Honcho AI integration to provide spending insights and budget recommendations based on transaction patterns.

Built With

• connect-pg-simple
• date-fns
• drizzle-orm
• esbuild
• express-session
• express.js
• flowglad-api
• honcho-ai
• lucide-react
• nanoid
• neon
• node.js
• passport.js
• postgresql
• radix-ui
• react-18
• react-hook-form
• replit
• shadcn/ui
• tailwind-css
• tanstack-query-v5
• typescript
• vite
• wouter
• zod