Up Banking API Skill

An AI agent skill for interacting with the Up Banking API. Query accounts, transactions, categories, and tags.

Installation

Via skills.sh

npx skills add tcn33/up-banking-api-skill

Manual installation

Clone the repo: git clone https://github.com/tcn33/up-banking-api-skill.git ~/.claude/skills/up-api.

Alternative: Download the ZIP and extract to ~/.claude/skills/up-api.

Setup

1. Get your API token

1. Visit api.up.com.au
2. Scan the QR code with your Up app
3. Copy the generated token

2. Set the environment variable

export UP_API_TOKEN="up:yeah:xxxxxxxxxxxxx"

Add this to your shell profile (~/.zshrc or ~/.bashrc) for persistence.

3. Verify setup

python3 scripts/up_api.py ping

You should see a status message with an emoji and UUID.

Usage

All listing commands support --json for raw JSON output.

Accounts

# List all accounts
python3 scripts/up_api.py accounts

# Filter by type (saver, transactional, home_loan)
python3 scripts/up_api.py accounts --type saver

# Filter by ownership (individual, joint)
python3 scripts/up_api.py accounts --ownership joint

# Get specific account details
python3 scripts/up_api.py account <account_id>

Transactions

# Recent transactions (default 30)
python3 scripts/up_api.py transactions

# Limit results
python3 scripts/up_api.py transactions --limit 10

# Filter by date range (RFC-3339 format required)
python3 scripts/up_api.py transactions --since 2024-01-01T00:00:00Z --until 2024-02-01T00:00:00Z

# Filter by status
python3 scripts/up_api.py transactions --status settled
python3 scripts/up_api.py transactions --status held

# Filter by category
python3 scripts/up_api.py transactions --category groceries

# Filter by tag
python3 scripts/up_api.py transactions --tag "Coffee"

# Filter by account
python3 scripts/up_api.py transactions --account <account_id>

# Fetch all pages (for large date ranges)
python3 scripts/up_api.py transactions --all --since 2024-01-01T00:00:00Z

# Get specific transaction details
python3 scripts/up_api.py transaction <transaction_id>

Categories

# List all categories
python3 scripts/up_api.py categories

# List subcategories of a parent
python3 scripts/up_api.py categories --parent good-life

# Set transaction category
python3 scripts/up_api.py categorize <transaction_id> <category_id>

# Clear transaction category
python3 scripts/up_api.py categorize <transaction_id> none

Tags

# List all tags
python3 scripts/up_api.py tags

# Add tags to a transaction
python3 scripts/up_api.py tag <transaction_id> "Coffee" "Work"

# Remove tags from a transaction
python3 scripts/up_api.py untag <transaction_id> "Coffee"

Webhooks

# List webhooks
python3 scripts/up_api.py webhooks

# Create a webhook
python3 scripts/up_api.py webhook-create https://example.com/hook "My webhook"

# Test a webhook
python3 scripts/up_api.py webhook-ping <webhook_id>

# View webhook delivery logs
python3 scripts/up_api.py webhook-logs <webhook_id>

# Delete a webhook
python3 scripts/up_api.py webhook-delete <webhook_id>

JSON Output

Use --json with any listing command to get raw JSON for processing:

# Get all January transactions as JSON
python3 scripts/up_api.py transactions --json --all \
  --since 2024-01-01T00:00:00Z \
  --until 2024-02-01T00:00:00Z

# Pipe to jq for analysis
python3 scripts/up_api.py transactions --json --category groceries | \
  jq '[.[].attributes.amount.value | tonumber] | add'

Date Format

The Up API requires RFC-3339 format for date filters:

• 2024-01-15T00:00:00Z (UTC)
• 2024-01-15T00:00:00+11:00 (with timezone offset)

Simple date strings like 2024-01-15 will not work.

Files

• scripts/up_api.py - CLI tool
• skill.md - Skill definition for AI agents
• references/api-reference.md - API endpoint reference

Resources

• Up API Documentation
• Up API GitHub