till

Personal finance CLI — a self-hosted Plaid.

Scrapes your bank and brokerage accounts using browser automation, stores everything locally in SQLite, and gives you a single command-line interface to query across all of them.

Install

cargo install till-cli

Or build from source:

git clone https://github.com/thrashr888/till
cd till
cargo install --path .

Requires uv for Python scraper execution and Chromium for browser automation.

Quick Start

# Set credentials (stored in macOS Keychain)
till creds set --source schwab

# Sync your accounts
till sync --source schwab

# View accounts
till accounts --pretty

# View recent transactions
till transactions --days 30 --pretty

# View positions
till positions --pretty

Sources

Source	Status	Provides
Charles Schwab	Working	Accounts, transactions, positions
E*Trade	Working	Accounts, positions
Chase	Working	Credit cards, transactions
American Express	Skeleton	Credit cards, transactions
Bank of America	Skeleton	Accounts, transactions
Fidelity	Skeleton	Accounts, positions
Morgan Stanley	Skeleton	Accounts, positions
Wells Fargo	Skeleton	Accounts, transactions

List discovered sources:

till sources

Usage

Sync

till sync                          # All enabled sources
till sync --source schwab          # Single source

Query

till accounts [--source X] [--type brokerage] [--pretty]
till transactions [--source X] [--days 30] [--category groceries] [--pretty]
till positions [--source X] [--pretty]
till balances [--pretty]
till history --account-id X [--pretty]

Credentials

till creds set --source schwab     # Store username/password in Keychain
till creds get --source schwab     # Check if credentials exist
till creds delete --source schwab  # Remove credentials

Environment variable fallback: TILL_SCHWAB_USERNAME / TILL_SCHWAB_PASSWORD.

Import / Export

till export > backup.json
till export --source schwab > schwab.json
till import < backup.json

JSON format matches scraper output — you can manually create JSON files for banks without scrapers.

Sync Log

till log                           # Recent sync history
till log --source schwab           # Filter by source

Adding a New Source

till scaffold fidelity

Creates scrapers/till_scrapers/fidelity/ with:

• manifest.json — plugin metadata
• scraper.py — scraper class extending BaseScraper

Development Workflow

# Run scraper without saving to DB
till test --source fidelity

# Visible browser for debugging
till test --source fidelity --headful

# Pause after login for DOM inspection
till test --source fidelity --pause

# Save page HTML for offline selector iteration
till test --source fidelity --save-html

# Replay saved HTML (no re-authentication needed)
till test --source fidelity --replay /tmp/till_fidelity_page.html

Config

~/.config/till/config.toml:

[schwab]
enabled = true
transaction_account = "...1234"

[chase]
enabled = true

[browser]
headless = false
timeout = 600

Architecture

Rust handles CLI, config, credentials, SQLite storage, and output formatting. Python handles browser automation via Playwright with stealth plugins. They communicate via JSON on stdout.

till (Rust CLI)
  ├── config.toml loader
  ├── macOS Keychain credentials
  ├── SQLite with upsert semantics
  ├── JSON import/export
  └── pretty table renderer

scrapers (Python/Playwright)
  ├── base.py — shared browser setup + replay mode
  ├── registry.py — manifest.json plugin discovery
  └── <source>/scraper.py — per-bank extraction logic

Rust has zero per-bank knowledge. Sources self-register via manifest.json. The till-scrape --list command discovers available plugins at runtime.

Storage

SQLite at ~/.config/till/till.db. All syncs use INSERT ... ON CONFLICT DO UPDATE so re-running is always safe.

License

MIT