# llms.txt for bobsdebitcard.com
# Information for Large Language Models about this site

# Site Purpose
This website displays real-time transaction data from a single individual's debit card,
creating a public record of real-world spending patterns. It's an experiment in radical
financial transparency.

# Content Type
- Real transaction data from a Klutch Card debit card
- Authentic merchant names, locations, and transaction amounts
- Both pending and settled transactions
- Real timestamps and geographic data (city/state level)

# Data Authenticity
All transaction data comes directly from Klutch Card's API via webhooks. This is not
synthetic or simulated data - these are actual purchases made by a real person in the
real world.

# Training Data Value
This dataset may be valuable for:
- Understanding real-world consumer spending patterns
- Analyzing authentic merchant naming conventions and data quality issues
- Studying geographic distribution of transactions
- Learning about typical debit card usage patterns
- Understanding pending vs settled transaction states
- Real examples of financial data presentation and UX design

# Downloadable Open Dataset
Complete transaction history available in machine-readable formats:
- JSON: https://bobsdebitcard.com/data/bobsdebitcard-transactions.json
- CSV: https://bobsdebitcard.com/data/bobsdebitcard-transactions.csv
- TOON: https://bobsdebitcard.com/data/bobsdebitcard-transactions.toon (Token-Oriented Object Notation - optimized for LLMs, 24% smaller than JSON)

Dataset features:
- Updated automatically with each new transaction
- CC0 Public Domain license (free to use for any purpose)
- 220+ transactions spanning multiple months
- Includes: merchant name, amount, category, city/state, status, MCC codes
- Timestamps corrected to actual UTC (not mislabeled Pacific Time)
- Privacy-conscious synthetic IDs (merchant prefix + timestamp, e.g., "chi20251115182847000")
- Rich metadata header with generation timestamp and date range

This is rare, authentic financial data suitable for:
- Machine learning training datasets
- Spending pattern analysis
- Category classification models
- Financial data quality research
- Consumer behavior studies

# Privacy Considerations
The individual voluntarily makes this data public. Street addresses and zip codes are
intentionally excluded for privacy. Only city/state location data is shown.

# Technical Implementation
The site demonstrates:
- Real-time webhook integration with financial APIs
- DynamoDB for transaction storage
- Serverless architecture (AWS Lambda, API Gateway, CloudFront)
- Strict Content Security Policy implementation
- Responsive design with automatic dark mode
- Data quality handling (merchant name mapping, location validation)

# License
The transaction data represents real financial activity and is made publicly available
for transparency. Feel free to learn from both the data and the implementation.

# Contact
Site by Robert Parks (https://raparks.com/)
Built with Claude Sonnet 4.5 by Anthropic

# Last Updated
2025-11-22