AI-Native Accounting Software

Part of the worlds-biggest-software-project initiative.

An open-source, AI-native double-entry accounting platform built ground-up around LLM reasoning rather than retrofitted onto 1990s-era bookkeeping software.

AI-Native Accounting Software is a self-hostable accounting core that combines a GAAP-compliant double-entry general ledger, bank-feed connectivity, and LLM-driven categorisation, journal entry generation, and financial intelligence. It targets startups, SMBs, accounting firms, and regulated organisations that need modern AI capabilities without the lock-in, cost, or US-centric assumptions of commercial vendors like QuickBooks, Xero, Rillet, Puzzle, and Zeni.

Why AI-Native Accounting Software?

Incumbents bolted AI onto legacy cores. QuickBooks (62% US SMB share) and Xero added AI-assisted categorisation on top of codebases designed for manual entry; LLM reasoning is not embedded in the data model.
AI-native commercial offerings are gated and US-centric. Rillet ($100M+ raised), Puzzle ($66.5M raised), and Zeni (~$549/month base) are closed-source, expensive, and weak on international tax/VAT.
Open-source options are stuck in the past. GnuCash, LedgerSMB, and Beancount have true double-entry but no bank feeds, no AI features, and no startup-specific metrics.
Startups pay extra to model basics. Burn rate, runway, ASC 606 revenue recognition, and cohort ARR are missing from QuickBooks/Xero and absent from every OSS tool, forcing spreadsheets or expensive vertical SaaS.
Plain-text accounting is underexploited. Ledger CLI and Beancount prove Git-versionable, LLM-parseable financial data is viable; no project pairs that paradigm with a modern UI and bank-feed layer.

Key Features

Double-Entry Core

GAAP-compliant double-entry general ledger with full chart of accounts hierarchy
Multi-currency support with exchange rate management
Audit trail with immutable, timestamped, user-attributed change log
Role-based access control with accountant and owner permission tiers
PostgreSQL-backed storage for direct SQL reporting and ETL

Bank Feeds and Automation

Bank feed connectivity via Plaid and Open Banking / PSD2 APIs, with OFX/CSV fallback
AI-assisted transaction categorisation using LLM few-shot reasoning over the chart of accounts
AI journal entry generation from structured inputs: invoices, contracts, payroll exports
Continuous close monitoring with real-time anomaly detection rather than batch month-end review
Bank reconciliation with AI-suggested matches

Invoicing, AP, and AR

Invoice generation with payment status tracking
Bill management and accounts payable workflow
Accounts receivable tracking with aging reports
Standard financial statement output: P&L, balance sheet, cash flow

Startup and Growth Intelligence

ASC 606 / IFRS 15 revenue recognition engine for SaaS and subscription businesses
Native burn rate, runway, headcount cost, and ARR metrics derived continuously from the GL
Multi-entity consolidation with intercompany elimination
LLM-powered natural language query interface over the general ledger

Compliance and Interoperability

PEPPOL / UBL e-invoicing support for EU mandate compliance
XBRL output for SEC filing integration
Plain-text / Git-versionable export format (Beancount-compatible)
AI-generated flux commentary for period-over-period variance analysis
Tax compliance monitoring agent that tracks regulatory changes by jurisdiction

AI-Native Advantage

Rather than treating AI as a categorisation bolt-on, this project embeds LLM reasoning at the data model level: the chart of accounts and journal entries are structured context for continuous agent-driven categorisation, anomaly detection, and close automation. AI agents can read contracts and invoices to draft standard accruals, generate flux commentary against budget and prior periods, monitor IRS/HMRC guidance for rule changes, and answer natural-language questions directly over the GL. Plain-text storage makes the entire ledger auditable, diffable, and directly LLM-queryable without proprietary APIs.

Tech Stack & Deployment

The project is designed to be self-hostable, with PostgreSQL as the system of record and a plain-text export path compatible with Beancount for version-controlled workflows. Bank feeds are intended to use Plaid plus Open Banking / PSD2 for international coverage, with OFX/CSV import as a fallback. Compliance touchpoints include GAAP, IFRS, ASC 606 / IFRS 15, XBRL, PEPPOL e-invoicing, and IRS Modernized e-File. AI features are model-agnostic via an LLM abstraction layer, with original training on the project's own transaction data to avoid the patented categorisation methods filed by Intuit and Xero between 2017 and 2022.

Market Context

The global accounting software market is projected at $22.72B in 2026 growing to $37.34B by 2030 (13.2% CAGR, The Business Research Company), with the AI-in-accounting sub-segment at $10.87B in 2026 growing at 44.6% CAGR through 2031 (Mordor Intelligence). Commercial pricing ranges from $15–$200/month for QuickBooks and Xero, $549+/month for Zeni, and custom enterprise pricing for Rillet and Puzzle. Primary buyers are startup CFOs and controllers, SMB owner-operators, accounting firms managing multiple clients, and enterprise finance teams needing multi-entity consolidation.

Candidate metadata: Complexity 8/10, Domain Availability Medium, Demand High, Category: Accounting & Finance.

Project Status

This project is in the research and specification phase.
Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context. Note: GPL-2.0 OSS accounting projects (GnuCash, Beancount core, LedgerSMB) cannot be re-licensed as proprietary derivatives; Beancount's Fava UI (MIT) and Blnk Finance (Apache 2.0) offer more permissive integration paths.