QIF Agent

QIF Agent is an AI-powered financial assistant that lets you chat with your own QIF (Quicken Interchange Format) data. It parses your QIF files, stores transactions in a SQLite database, and uses an LLM (via Ollama) to answer natural language questions about your finances. The project includes a FastAPI backend and a Streamlit-based web UI.

Start the Streamlib web UI, then ask it queries like

• List all transactions from 2021 where category like Util or like Electric
• What the sum total for all of 2018 where the category like Dues?

The Agent creates a SQLite database called "transactions.db". A table "transactions" is created and has the following fields:

• date
• payee
• category
• memo
• amount

Features

• QIF Parsing: Indexes and parses QIF files into a structured SQLite database.
• Natural Language Chat: Ask questions about your transactions, totals, trends, and more.
• Web UI: Simple chat interface built with Streamlit.
• Dockerized: Easy to run with Docker Compose.

Project Structure

.
├── app/                # Backend FastAPI app and QIF indexer
│   ├── main.py
│   └── qif_indexer.py
├── db/                 # SQLite database storage (created at runtime)
│   └── transactions.db
├── qifs/               # Place your QIF files here
│   └── pb2024.qif
├── ui/                 # Streamlit UI
│   ├── qif_chat.py
│   ├── requirements.txt
│   └── Dockerfile
├── requirements.txt    # Backend Python dependencies
├── Dockerfile          # Backend Dockerfile
├── docker-compose.yml  # Multi-service orchestration
└── .gitignore

Prerequisites

• Docker and Docker Compose
• Ollama running locally or accessible (for LLM inference)
• At least one .qif file in the qifs/ directory

Quick Start

1. Start Ollama

   Make sure Ollama is running and the phi4-mini:3.8b model is available:

   ollama run phi4-mini:3.8b

2. Add Your QIF Files

   Place your .qif files in the qifs/ directory.

3. Build and Run with Docker Compose

   docker-compose up --build

   • The FastAPI backend will be available at http://localhost:8000
   • The Streamlit UI will be available at http://localhost:8501

4. Chat with Your QIF Agent

   Open http://localhost:8501 in your browser and start asking questions about your finances!

API Endpoints

• POST /chat — Ask a question about your transactions (used by the UI)
• GET /transactions/{year} — List transactions for a given year
• GET /transactions/count/{year} — Count transactions for a given year
• GET /count — Total number of transactions
• GET /health — Health check

Environment Variables

Set in docker-compose.yml:

• QIF_DIR — Path to QIF files (default: /qifs)
• DB_PATH — Path to SQLite database (default: /db/transactions.db)
• OLLAMA_URL — URL for Ollama server (default: http://host.docker.internal:11434)

Development

• Backend code: app/main.py, app/qif_indexer.py
• UI code: ui/qif_chat.py

License

MIT License. See LICENSE if present.

---

Built with FastAPI, Streamlit, LangChain, and Ollama running LLM Ph4-mini:3.8b.