Categories Configuration

The file config/categories.yaml defines spending categories, auto-categorisation rules, and tag rules. Copy from the example:

cp config/categories.yaml.example config/categories.yaml

The file has three sections: categories, rules, and tag_rules.

Categories

Define all spending and income categories:

categories:
  # Income
  - name: Salary
    is_income: true
  - name: Interest Income
    is_income: true

  # Expenses with optional monthly budgets
  - name: Groceries
    budget_monthly: 600
  - name: Eating Out
    budget_monthly: 300
  - name: Utilities
    budget_monthly: 300

  # Business (ATO deductible)
  - name: "Business: Hosting & Infrastructure"
  - name: "Business: Software & Subscriptions"

  # Financial
  - name: Transfers
  - name: Uncategorized

Category fields

Field	Required	Description
name	Yes	Unique category name
is_income	No	Set true for income categories
budget_monthly	No	Monthly budget amount for the budget tracker

Naming convention: Categories starting with Business: are automatically allocated at 100% to your business in the split engine. Categories starting with Rental: are for rental property expenses.

Rules

Rules assign categories to transactions automatically. They are matched top-to-bottom and first match wins.

Rule types

Source-based rules match all transactions from a source:

rules:
  - type: source
    source_type: airbnb
    category: Airbnb Income

Regex rules match against the transaction description:

  - type: regex
    pattern: "WOOLWORTHS|COLES |ALDI |IGA |HARRIS FARM|COSTCO"
    category: Groceries

Patterns are matched case-insensitively against the uppercase description.

Rule ordering matters

Rules are evaluated top-to-bottom. Put specific patterns before generic ones:

rules:
  # Specific: credit card interest (before mortgage interest)
  - type: regex
    pattern: "INTEREST CHARGED"
    category: Fees & Charges

  # Generic: mortgage interest
  - type: regex
    pattern: "INTEREST CHARGE"
    category: Mortgage

Business rules should go near the top so they match before personal categories:

rules:
  # Business hosting (matched first)
  - type: regex
    pattern: "NAMECHEAP|DIGITAL ?OCEAN|CLOUDFLARE|AWS"
    category: "Business: Hosting & Infrastructure"

  # Business software
  - type: regex
    pattern: "GITHUB|FIGMA|ADOBE|OPENAI|ANTHROPIC"
    category: "Business: Software & Subscriptions"

  # Personal shopping (matched later)
  - type: regex
    pattern: "AMAZON|EBAY"
    category: Shopping

Adding a new merchant

To add a merchant, append a regex rule. If the merchant name appears differently across banks, use alternation:

  - type: regex
    pattern: "UBER ?EATS|DOORDASH|MENULOG|DELIVEROO"
    category: Eating Out

Tips for writing patterns:

• Use | to match multiple merchants in one rule
• Use \\b for word boundaries when a pattern is too broad
• Use ^ to anchor at the start of the description
• Use \\d to match digits (e.g. 7-ELEVEN \\d)
• Use \\. to match a literal dot (e.g. APPLE\\.COM)
• Spaces matter: "COLES " (with trailing space) avoids matching “COLESLAW”

Confidence levels

The categoriser assigns a confidence score:

Match type	Confidence
Source-based rule	1.0
Exact match	0.9
Regex match	0.8
Learned (from manual override)	0.7
Uncategorized	0.0

Tag rules

Tags are orthogonal to categories — a transaction has one category but can have many tags. Tags enable sub-classification for reporting.

tag_rules:
  # Transport sub-types
  - tag: flight
    pattern: "JETSTAR|QANTAS|VIRGIN AUSTRALIA"
  - tag: taxi
    pattern: "UBER TRIP|UBER AUSTRALIA|TAXI"
  - tag: train
    pattern: "OPAL|TRANSPORTFORNSW"

  # Insurance sub-types
  - tag: health-insurance
    pattern: "MEDIBANK|AMBULANCE"

  # Business sub-types (for ATO detail)
  - tag: biz-hosting
    pattern: "NAMECHEAP|DIGITAL ?OCEAN|CLOUDFLARE|AWS"
  - tag: biz-software
    pattern: "GITHUB|FIGMA|ADOBE"

  # Utility sub-types
  - tag: internet
    pattern: "TPG"
  - tag: mobile
    pattern: "OPTUS|VODAFONE"

Tags are used by the business split engine to determine which portion of a bill is business-deductible (e.g. 50% of the internet-tagged utility bill).

Tag rule fields

Field	Required	Description
tag	Yes	Tag name (lowercase, hyphenated)
pattern	Yes	Regex pattern matched against description
source_type	No	Only match transactions from this source

Learned rules

When you manually change a transaction’s category in the dashboard, the system stores the description-to-category mapping in the category_rules_learned database table. Next time a transaction with the same description appears, it will be auto-categorised with confidence 0.7.

After editing

Run ledger init to reload categories into the database:

ledger init

To re-categorise existing transactions with new rules, re-ingest the archived files or use the dashboard to update individual transactions.