AI Bank Statement Analyzer

Given an arbitrary bank statement PDF, utilize AI to analyze and provide insights and generate a report.

Example Output

Architecture

Flow Outline

Stack

Database Modeling

Notes

Setup (Local Dev)

Ensure you have ./frontend/.env and ./backend/.env setup correctly.

From project root directory:

• yarn install
• yarn prisma:generate

Then stand up the services for the backend:

• docker compose up postgres redis

Now migrate / seed the db:

• yarn setup:backend

Now you can run the entire app with

• yarn dev

Frontend is on http://localhost:5173

Backend is on http://localhost:3000

Setup (Docker)

Ensure you have ./frontend/.env and ./backend/.env setup correctly.

From project root directory:

• docker compose up

Frontend is on http://localhost:5173

Backend is on http://localhost:3000

Additional Scripts

• yarn prepare - setup for husky precommit hooks

Backend

• yarn db:explore (Spins up browser to view database using prisma at http://localhost:5555)
• yarn db:reset (nuke it)

Backend .env

The only thing to change when using the docker setup is DATABASE_URL (uncomment the one with localhost if working locally):

• DATABASE_URL="postgresql://mr_krabs:secretformula@localhost:5432/krusty_krab?schema=public

And uncomment REDIS_HOST and REDIS_PORT as well if running with docker.

Be sure to add your OPENAI_API_KEY regardless.

NODE_ENV="development"

# Express Server
PORT=3000
CORS_WHITELIST="http://localhost:3000, http://localhost:5173"

# Database
POSTGRES_USER="mr_krabs"
POSTGRES_PASSWORD="secretformula"
POSTGRES_HOST="postgres"
POSTGRES_PORT=5432
POSTGRES_DB="krusty_krab"

DATABASE_URL="postgresql://mr_krabs:secretformula@localhost:5432/krusty_krab?schema=public"
# DATABASE_URL="postgresql://mr_krabs:secretformula@postgres:5432/krusty_krab?schema=public"

# Redis for Queue
# REDIS_HOST=redis
# REDIS_PORT=6379

# Local Storage
FILE_UPLOAD_MAX_SIZE_BYTES=52428800

# OpenAI
OPENAI_API_KEY="sk-proj-ABCD1234"