AI Money

This is a personal finance expense tracker. This works on top of transactions in beancount file format. The app converts an uploaded csv of credit card statement into a beacnount file and categorizes the expenses with an AI agent.

Features

• Categorize transactions using AI

Setup

Backend

• Install the python version specified in pyproject.toml
• cd backend
• pip install poetry
• poetry install

Frontend

This is a Next.js app bootstrapped with create-next-app.

Install

• Node.js (version 22.x or later recommended)
• npm (comes with Node.js)
• cd frontend
• npm install

Run

Backend

To build and run the application:

1. Install dependencies:

cd backend
poetry shell
poetry install

2. Run the server:

ANTHROPIC_API_KEY=<your_key> poetry run uvicorn app:app --reload

Build and run the application using Docker:

Build the Docker image:

docker build -t ai-money/backend .

Test locally:

docker run -p 8000:8000 -e ANTHROPIC_API_KEY=your_key ai-money/backend

AWS Deployment

docker build --platform=linux/amd64 -t ai-money/backend:`git log -n 1 --format="%H"` .
docker tag ai-money/backend:`git log -n 1 --format="%H"` 867344451303.dkr.ecr.us-west-2.amazonaws.com/ai-money/backend:`git log -n 1 --format="%H"`
aws ecr get-login-password --region us-west-2 | docker login --username AWS --password-stdin 867344451303.dkr.ecr.us-west-2.amazonaws.com
docker push 867344451303.dkr.ecr.us-west-2.amazonaws.com/ai-money/backend:`git log -n 1 --format="%H"`

Frontend

To build and run the application:

cd frontend
npm run dev

And visit localhost:3000

Build and run the application using Docker:

# Build the Docker image
docker build -t ai-money/frontend .

# Run the container
docker run -p 3000:3000 -e BACKEND_HOST=localhost:8000 ai-money/frontend

AWS Deployment

docker build --platform=linux/amd64 -t ai-money/frontend:`git log -n 1 --format="%H"` .
docker tag ai-money/frontend:`git log -n 1 --format="%H"`  867344451303.dkr.ecr.us-west-2.amazonaws.com/ai-money/frontend:`git log -n 1 --format="%H"`
aws ecr get-login-password --region us-west-2 | docker login --username AWS --password-stdin 867344451303.dkr.ecr.us-west-2.amazonaws.com
docker push 867344451303.dkr.ecr.us-west-2.amazonaws.com/ai-money/frontend:`git log -n 1 --format="%H"`

Sample CC statement

A sample statement is available in backend/statements. Currently, the application supports only this statement format. To use this application, ensure your statements are converted to this CSV format.

Dashboard

Paisa is a dashboard tool to visualize the transactions. To visualiase the transactions you can use that.