Skip to content
OpenLedger

A reliable, deterministic ledger for your AI.

OpenLedger is the harness that keeps your AI from hallucinating your money. It turns scattered bank and credit-card statements into a source of truth you can build any financial AI agent on.

npm install -g @aquartier/openledger

Then paste each step into your AI. See the install guide

Works with the AI you already use

  • Claude Code
  • Codex
  • Cursor
  • Grok Build
  • PIPI
  • OpenCode
  • OpenClaw
  • HHermes

Why OpenLedger

A solid financial ground for your AI to work on

Left alone, AI juggles numbers in its head, and sooner or later it will hallucinate. With OpenLedger, your money is kept in a reliable double-entry ledger instead, so every figure your AI quotes is looked up, never made up.

Your own private Plaid for your AI

Scattered statements go in, a reliable ledger comes out. The paperwork your bank already sends becomes the source of truth your AI works from, a private data feed you own end to end.

Local first, PII-masked, security built in

Everything stays on your machine as local bookkeeping. There’s no cloud, no bank login, no external server storing your data. Your financial history lives in a single local database, and is scrubbed (PII-masked) before anything reaches to external.

Install OpenLedger

Works seamlessly withClaude CodeCodexCursorGrok BuildPIPIOpenCodeOpenClawHHermesClaude Code

  1. Install the CLI

    npm install -g @aquartier/openledger

    Installs the oled command.

  2. Run OCR locally (optional)

    Download Typhoon-OCR-1.5-2B in LM Studio and start its local server, then hand this to your AI.

    Configure OpenLedger to use my local OCR at http://localhost:1234/v1 (model typhoon-ocr1.5-2b), then run oled doctor to confirm.

    Only scans and photos need OCR; PDFs with a text layer are read directly. Set it up if you expect a lot of payment slip photos in your project.

  3. Add the skill

    npx skills add phureewat29/openledger

    Or run oled setup. It writes the skill to the skills directory most agents read.

  4. Drop in your statements

    oled open

    Opens the data folder. Drag in your bank and credit-card statements, as PDFs or payment slip photos; subfolders are fine.

  5. Ingest them

    Ingest my credit cards and bank statements. Then show me anything you weren't sure about, and let's resolve it.

    Paste this into your AI. It will commit every transaction it can find to the secured ledger.

  6. Ask your first question

    What's my top spending?

    Congratulations! You now have a deterministic ledger to build any financial AI agent on; start with the example prompts below.

What AI agent you can build

Every agent below is a prompt away. One ledger underneath, so the numbers agree across all of them.

Spending Tracker

OCR required

Prompt:

“Build me a spending tracker that runs on my payment slip photos: read each slip, log the merchant, amount, and date, and update my charts of daily spend, top merchants, and category trends automatically.”

A private YNAB that already knows where your money goes.

Personal CFO

Prompt:

“Be my CFO. Criticize with real numbers, lead with what they mean, and have an opinion: what to cut, keep, or automate and why. Flag anything unusual before I ask, and end every answer with a next step.”

An advisor that knows you best and tells no one.

Money Visualizer

Prompt:

“Build me a money visualizer: a Sankey flow diagram of where my money went, with what-if sliders for income, spending, and saving, each projected from my real history so I can see where I end up.”

Every dollar flow drawn out, with projected what-ifs.

And your own.

Describe it. It ships tonight.

Whatever you imagine, running on deterministic numbers.

Under the hood

The model reads your statements; the ledger checks the math. Every posting balances, or it doesn’t post.

  • Double-entry ledger

    Every transaction debits one account and credits another, so the same input always produces the same books.

  • Local SQLite

    One file in ~/.oled with 0600 permissions. Your machine is the database.

  • PII redaction

    Names, account numbers, and free text are masked before anything goes to your AI provider. The ledger file itself never leaves.

  • Reference datasets

    Built-in institution data for the US, Japan, Thailand, and China, so your AI recognizes the bank behind every statement.

  • Questions, not guesses

    When a line is ambiguous, OpenLedger surfaces it for you to resolve instead of letting the model pick.

100%

of your ledger stays in one local file

Zero

bank logins or credentials

100%

of challenges pass with small model

1/9

token usage vs without harness

Benchmarked with small models

The harness carries the bookkeeping, so the model doesn’t have to be huge. A small model passes every challenge. Full benchmarking results in openledger-eval.

  • challenge:statement-rows-posted

    PASS

  • challenge:record-tasks

    PASS

  • challenge:ledger-queries

    PASS

  • total:cost

    $0.47 per trial

  • total:token-usage

    651K in / 56.6K out