No user-submitted related servers found.
Scored across 72 tools
Resources are generally separable (customer/product/account/tax/invoice/estimate), but the set contains many near-parallel or easily confused pairs: wave_list_transactions vs wave_list_ledger_transactions, wave_get_transaction vs wave_get_ledger_transaction, wave_create_transaction vs wave_create_transactions, wave_send_invoice vs wave_mark_invoice_sent, and wave_list_payments vs the transaction listers. The long descriptions do resolve most conflicts (live web-app data vs imported Data Export snapshot), but an agent must read carefully to avoid picking the wrong twin, and wave_graphql_query overlaps with most read tools.
Names follow a steady wave_verb_noun pattern with a shared verb vocabulary (list/get/create/update/delete/archive/approve/send/mark), and the invoice/estimate/transaction families mirror each other predictably. Minor deviations exist: wave_create_transactions (plural batch) next to wave_create_transaction, 'record' vs 'create' for payments/deposits, and one-off verbs like run_report, import_ledger, graphql_query.
72 tools is far beyond a comfortable surface for an agent, even for a broad accounting domain. Much of the bulk comes from duplicated CRUD families (live transaction API, imported-ledger reader, and parallel invoice/estimate payment lifecycles), which inflates the count without adding distinct capability.
Coverage is unusually thorough: customers, products, accounts, sales taxes, invoices, estimates, invoice payments, estimate deposits, transactions, ledger import, and a large report suite, each with read/create/update/delete-or-archive paths. Remaining gaps (vendor writes, bills, un-archive, payroll reports) are explicitly documented as Wave API limitations rather than omissions, and escape hatches (wave_graphql_query, ledger import) cover the rest.
