refund-radar

A local-first, privacy-first bank statement auditor that detects recurring charges, flags suspicious transactions, and drafts ready-to-send refund requests.

What This Skill Does

1. •Parses bank/card statement exports (CSV or pasted text)
2. •Detects recurring subscription charges
3. •Flags unexpected, duplicate, and fee-like transactions
4. •Generates an interactive HTML audit report
5. •Creates ready-to-send refund request templates
6. •Learns from your decisions to improve future scans

Trigger Phrases

• •"Scan my bank statement for refunds"
• •"Analyze my credit card transactions"
• •"Find recurring charges in my statement"
• •"Check for duplicate or suspicious charges"
• •"Help me dispute a charge"
• •"Generate a refund request"

How To Use

Step 1: Get Your Statement

Export transactions from your bank as CSV, or copy-paste them as text.

Step 2: Analyze

# From CSV file
python -m refund_radar analyze --csv ~/Downloads/statement.csv --month 2026-01

# From pasted text (stdin)
python -m refund_radar analyze --stdin --month 2026-01 --default-currency CHF

Step 3: Review Flags

The tool will:

• •Show recurring subscriptions with estimated next charge
• •Flag unexpected charges (new merchants, spikes, duplicates)
• •Ask clarifying questions in batches of 5-10

Step 4: Get Your Report

Output files are saved to:

• •~/.refund_radar/reports/YYYY-MM.html (interactive report)
• •~/.refund_radar/reports/YYYY-MM.json (raw data)

Step 5: Send Refund Requests

Copy refund templates directly from the HTML report. Templates are pre-filled with:

• •Merchant name and date
• •Charge amount
• •Dispute reason
• •Your preferred tone (concise, firm, friendly)

CLI Reference

# Analyze a statement
python -m refund_radar analyze --csv path/to/file.csv --month 2026-01

# Analyze from stdin
python -m refund_radar analyze --stdin --month 2026-01 --default-currency CHF

# Mark a merchant as expected (wont flag in future)
python -m refund_radar mark-expected --merchant "AMZN Mktp"

# Mark a merchant as known recurring
python -m refund_radar mark-recurring --merchant "Spotify"

# List expected merchants
python -m refund_radar expected

# Reset all learned state
python -m refund_radar reset-state

# Export month data as JSON
python -m refund_radar export --month 2026-01 --out out.json

Files Written

Privacy Note

• •No network calls. Everything runs locally.
• •No external APIs. No Plaid, no cloud services.
• •Your data stays on your machine. Reports are local HTML files.
• •Privacy toggle in reports. Blur merchant names with one click.

Detection Logic

Recurring Charges

• •Same merchant >= 2 times in 90 days
• •Similar amounts (within 5% or 2.00)
• •Consistent cadence (weekly, monthly, yearly)
• •Known subscription keywords (Netflix, Spotify, etc.)

Unexpected Charges

• •New merchant: First time seeing this merchant AND amount > 30
• •Amount spike: > 1.8x your baseline for this merchant
• •Duplicate: Same merchant + amount within 2 days
• •Fee-like: Contains FEE, COMMISSION, ATM, OVERDRAFT keywords
• •Currency anomaly: Unusual currency or DCC indicator

Example Flow

$ python -m refund_radar analyze --csv bank.csv --month 2026-01

Parsed 47 transactions for 2026-01

RECURRING CHARGES (5)
  Netflix          $15.99/mo   next: ~Feb 15
  Spotify          $11.99/mo   next: ~Feb 03
  iCloud           $2.99/mo    next: ~Feb 01
  ChatGPT Plus     $20.00/mo   next: ~Feb 20
  Gym Membership   $49.00/mo   next: ~Feb 10

FLAGGED (3)
  [HIGH] DUPLICATE: Amazon $29.99 on Jan 12 and Jan 12
  [MED]  AMOUNT SPIKE: Uber $87.50 (baseline: $25)
  [LOW]  FEE: ATM Withdrawal Fee $3.50

Report saved: ~/.refund_radar/reports/2026-01.html

Improving Over Time

The tool remembers:

• •Merchants you mark as expected
• •Recurring subscriptions you confirm
• •Your merchant aliases (e.g., "AMZN" = "Amazon")

This means fewer false positives on future runs.