CLI reference

API keys look like: sl_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx. Keep them secret and treat them like passwords.

sheetlink items

JSON (default)

The default output. A structured JSON object is written to stdout, making it composable with any Unix tool. Nothing is written to disk unless you redirect the output.

Output shape, shortened (each transaction also carries Plaid’s other fields, such as authorized_date, location, website and logo_url). Amounts follow Plaid’s sign: money out is positive.

{
  "synced_at": "2026-04-09T08:00:00.000Z",
  "items": [
    {
      "item_id": "VBX93wmRY4Iy...",
      "accounts": [
        {
          "account_id": "acc_abc123",
          "name": "Chase Checking",
          "official_name": "Chase Total Checking",
          "mask": "4242",
          "type": "depository",
          "subtype": "checking",
          "current_balance": 2847.12,
          "available_balance": 2647.12,
          "institution": "Chase"
        }
      ],
      "transactions": [
        {
          "transaction_id": "txn_xyz789",
          "account_id": "acc_abc123",
          "account_name": "Chase Checking",
          "account_mask": "4242",
          "date": "2026-04-08",
          "description_raw": "WHOLE FOODS MARKET #123",
          "merchant_name": "Whole Foods Market",
          "amount": 47.23,
          "iso_currency_code": "USD",
          "pending": false,
          "payment_channel": "in store",
          "personal_finance_category": {
            "primary": "FOOD_AND_DRINK",
            "detailed": "FOOD_AND_DRINK_GROCERIES"
          },
          "source_institution": "Chase"
        }
      ],
      "tier": "max",
      "days_available": 730
    }
  ]
}

# Every merchant name
sheetlink sync | jq '[.items[].transactions[].merchant_name]'

# Sum all transaction amounts
sheetlink sync | jq '[.items[].transactions[].amount] | add'

# Filter by merchant
sheetlink sync | jq '.items[].transactions[] | select(.merchant_name == "Amazon")'

CSV

Writes a flat CSV file. Each sync overwrites the file: it’s a point-in-time snapshot, not an append log. The default output path is ./sheetlink-transactions.csv.

CSV columns, in order. Pass --slim for the older 14-column set.

sheetlink sync --output csv --file ~/finances.csv

PostgresMax

Upserts transactions and accounts into a Postgres database. Tables are created automatically on first run. Safe to run repeatedly: deduplication is handled by transaction_id and account_id.

sheetlink sync --output postgres://user:password@host:5432/dbname

Auto-created schema:

-- Transactions: the 34 Google Sheets columns, plus category (Plaid's legacy category list)
CREATE TABLE IF NOT EXISTS sheetlink_transactions (
  transaction_id           TEXT PRIMARY KEY,
  account_id               TEXT,
  persistent_account_id    TEXT,
  account_name             TEXT,
  account_mask             TEXT,
  date                     DATE NOT NULL,
  authorized_date          DATE,
  datetime                 TEXT,
  authorized_datetime      TEXT,
  description_raw          TEXT,
  merchant_name            TEXT,
  merchant_entity_id       TEXT,
  amount                   DECIMAL(10,2),
  iso_currency_code        TEXT,
  unofficial_currency_code TEXT,
  pending                  BOOLEAN,
  pending_transaction_id   TEXT,
  check_number             TEXT,
  category_primary         TEXT,
  category_detailed        TEXT,
  payment_channel          TEXT,
  transaction_type         TEXT,
  transaction_code         TEXT,
  location_address         TEXT,
  location_city            TEXT,
  location_region          TEXT,
  location_postal_code     TEXT,
  location_country         TEXT,
  location_lat             DECIMAL(10,7),
  location_lon             DECIMAL(10,7),
  website                  TEXT,
  logo_url                 TEXT,
  source_institution       TEXT,
  category                 TEXT,
  synced_at                TIMESTAMP DEFAULT NOW()
);

-- Accounts
CREATE TABLE IF NOT EXISTS sheetlink_accounts (
  account_id            TEXT PRIMARY KEY,
  persistent_account_id TEXT,
  name                  TEXT,
  official_name         TEXT,
  mask                  TEXT,
  type                  TEXT,
  subtype               TEXT,
  current_balance       DECIMAL(10,2),
  available_balance     DECIMAL(10,2),
  iso_currency_code     TEXT,
  institution           TEXT,
  last_synced_at        TIMESTAMP DEFAULT NOW()
);

Pass --slim to create the older 14-column transactions table instead (date, name, amount, category, account_id, account_name, account_mask, source_institution, pending, payment_channel, merchant_name, category_primary, category_detailed, transaction_id).

SQLiteMax

Same schema as Postgres, but written to a local SQLite file. No server required, good for local development, personal finance dashboards, or lightweight automation.

sheetlink sync --output sqlite:///Users/you/finance.db

The file is created if it doesn’t exist, with the same sheetlink_transactions and sheetlink_accounts tables and columns as Postgres, in SQLite’s types (dates are stored as text).

Everything after sqlite:// is the file path, so three slashes give a full path and two give one from the current folder (sqlite://finance.db). From CLI 0.6.1, ~ works for your home folder too (sqlite:///~/finance.db).

Using env vars in practice:

# Set once in your shell profile
export SHEETLINK_API_KEY=sl_your_key_here
export SHEETLINK_OUTPUT=postgres://user:pass@host/mydb

# Then sync is just:
sheetlink sync