Use case · Finance teams

Can an AI post to the books without breaking them?

Letting an AI write to the books is only safe if the books refuse bad entries on their own, and keep a record of every change.

Talk to AttriView on GitHub

Step by step

How it works for finance. Each step is a real tool your AI calls.

OpenLedgerOpenWatch
  1. 01

    Every posting is checked

    Two or more entry lines, debits equal to credits, positive whole-cent amounts. Anything else is refused with the reason.

    post_transaction
  2. 02

    Mistakes are reversed, not edited

    A correction posts a reversal. The original stays, and a reversal cannot itself be reversed.

    reverse_transaction
  3. 03

    Every write is audited

    The audit row is written in the same database transaction as the posting.

    get_audit_log
  4. 04

    Reports come from the same books

    Trial balance, profit and loss, and balance sheet, straight from the ledger.

    get_trial_balance

Example

What it looks like. Sample data; the tool names are real.

Your AIOpenLedger + OpenWatch

Post the March software invoice: $1,200 to software expense, paid from operating cash.

  1. list_accounts
  2. post_transaction

Posted: debit Software $1,200.00, credit Cash $1,200.00. It balances, and the audit row was written with it.

Limits

What it doesn't do yet. So you know before you try it.

Need one of these for your team? Building it is the work Attri does.

Talk to Attri
  • One currency. No tax, accounts receivable or payable, or period close yet.
  • No bank feeds. Transactions arrive through your AI.
  • MCP only: there is no REST API, and no per-agent permissions yet.
  • SQLite only. No login in the open-source version.
Other teams
LegalWhere did every fact in this brief come from?ComplianceWhich staff used AI on this work, and can we prove it?

Start with one tool. Connect it to your AI today.

Clone it from GitHub and add it to your MCP client. When your team needs it hosted, with login, talk to Attri.

View on GitHubTalk to Attri