WBSP
ProvidersDocsSign in

TallyCrux

A variant of Expense Management Platform.

View the interactive variant page →

Expense Management Platform

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

An AI-native, open-source expense management platform that pairs multi-modal receipt capture with natural-language policy enforcement and synthetic-fraud defence — without locking customers into a proprietary corporate card.

This project is a card-agnostic expense management platform for finance teams who want modern AI capabilities without surrendering their banking relationship. It targets the gap between legacy enterprise suites (SAP Concur, Emburse) and US-centric card-revenue platforms (Ramp, Brex), and aims to give SMB and mid-market companies — particularly outside the US — a serious open-source alternative.

Why Expense Management Platform?

  • Incumbents tie expense intelligence to their card. Ramp and Brex (now Capital One) deliver real-time policy enforcement only when customers adopt their corporate card; companies that cannot or will not switch card programmes are excluded from the AI-driven tier of the market.
  • Legacy suites are AI-bolt-on, not AI-native. SAP Concur leads on global compliance but runs a 1990s-era architecture with bolted-on AI, dated UX, and policy configuration that requires IT involvement; user satisfaction lags newer competitors.
  • Synthetic receipt fraud is a 2025-era threat with no open-source defence. The March 2025 OpenAI image-generation upgrade made high-quality fake receipts trivial to produce; current OCR-only validation is vulnerable, and no open-source platform addresses this gap.
  • Existing OSS options are functionally thin. ERPNext Expense Claims and Odoo Community offer no receipt OCR, no AI policy enforcement, no real-time card feeds, and no mobile-first capture — they are accounting modules, not expense platforms.
  • Europe and the SMB segment are underserved. Pleo addresses European SMBs but lacks enterprise policy depth; no GDPR-aware, VAT-reclaim-capable open-source alternative exists for EU companies.

Key Features

Receipt Capture and Data Extraction

  • Mobile receipt capture with AI-powered OCR extraction of merchant, amount, date, currency, and VAT
  • Multi-modal AI processing capable of handling damaged, handwritten, low-light, and non-standard receipt images
  • Email-inbox forwarding with automatic receipt-to-transaction pairing
  • Multi-currency expense handling using ISO 4217 currency codes
  • Automatic VAT field extraction for EU expense reporting

Policy Enforcement and Compliance

  • Configurable policy rules engine: per-category limits, merchant category restrictions, and per-diem rates
  • Natural-language policy rule engine that interprets plain-English policy statements via an LLM and enforces them without IT configuration
  • Real-time point-of-purchase policy enforcement with immediate employee and manager notification
  • Per diem / GSA rate integration for US government contractors
  • VAT reclaim tracking aligned with EU Directive 2008/9/EC and the 13th Directive
  • IRS Accountable Plan compliance reporting

Fraud Detection and Audit

  • Synthetic receipt fraud detection combining image metadata analysis (EXIF), merchant network verification, and geographic plausibility checks
  • Behavioural anomaly scoring against each employee's historical merchant-category patterns
  • Cross-employee pattern detection for collusion and duplicate-submission fraud
  • Per-expense audit trail covering submission, review, and payment
  • AI-driven flagging of high-risk expenses prior to approval

Card and Banking Integration

  • Card-agnostic transaction feeds via Open Banking APIs (Plaid, FDX, PSD2)
  • Automatic receipt-to-transaction matching for any Visa or Mastercard feed
  • Pluggable card adapters so customers retain their existing banking relationships
  • Real-time spend feed with AI-categorised GL coding pushed automatically to the connected ERP

Reimbursement, Spend Intelligence, and ERP Sync

  • Employee reimbursement via ACH with payroll integration
  • ERP synchronisation for GL posting (ERPNext, QuickBooks, Xero, NetSuite, Sage Intacct)
  • HRIS integrations for card provisioning linked to onboarding (Gusto, Rippling, BambooHR)
  • Duplicate subscription detection and preferred-vendor savings recommendations
  • Natural-language spend queries for ad-hoc analysis without custom reporting

AI-Native Advantage

Unlike incumbents that bolt AI onto rule-based architectures, this platform is structured around AI from the data layer up. Multi-modal vision-and-text models extract accurate data from receipts that defeat template OCR; an LLM-based policy engine accepts plain-English rules and enforces them at the point of swipe rather than at reimbursement; and a layered fraud-defence model — combining EXIF metadata analysis, merchant verification, and behavioural anomaly scoring — addresses the synthetic-receipt threat that emerged after the March 2025 generative-image upgrade. Spend intelligence is proactive rather than reactive: redundant subscriptions, sub-optimal vendors, and budget overruns are surfaced from real-time spend data before they compound.

Tech Stack & Deployment

The platform is designed to be self-hostable for companies that need data residency or GDPR-aligned deployment, with a managed-cloud option for SMBs. Card and bank connectivity is built on Open Banking standards (Plaid and FDX in the US, PSD2 in the EU) so customers are not forced onto a proprietary card. ERP and HRIS connectivity is delivered through pre-built connectors for QuickBooks, Xero, NetSuite, Sage Intacct, ERPNext, Gusto, Rippling, and BambooHR, with a REST API for custom integrations. Compliance scope includes IRS Accountable Plan rules, EU VAT reclaim directives, GDPR, CCPA, and PCI DSS for card-data handling.

Market Context

The global expense management software market is estimated at $8.48B–$9.29B in 2026 and projected to reach $12.86B–$15.86B by 2030–2032 at a 9.25%–11.1% CAGR (Mordor Intelligence; Fortune Business Insights). Commercial pricing ranges from $5/user/month (Expensify) to an estimated $8–$15/user/month for SAP Concur enterprise deployments, while Ramp and pre-acquisition Brex monetise via card interchange. Primary buyers are finance managers and controllers at 50–500-employee companies, CFOs at enterprises focused on fraud reduction and audit completeness, and procurement leads consolidating cards, expenses, and AP. Capital One's $5.15B acquisition of Brex (January 2026) and Ramp's $32B valuation indicate strong mainstream financial-institution interest in AI-driven corporate spend.

Quickstart (local development)

Prerequisites: Docker (Compose v2), Node.js 22 LTS, pnpm 9+.

pnpm install
cp .env.local.example .env.local   # fill CHANGE_ME values (JWT keys: see the
                                   # generation one-liner inside the file)
pnpm dev:up                        # starts postgres/redis/minio, runs
                                   # migrations and the idempotent seed
pnpm --filter @expense-platform/api dev   # API on :3001
pnpm --filter @expense-platform/web dev   # web on :3000

Verify: curl localhost:3001/health reports every backing service connected; curl localhost:3001/docs/json serves the generated OpenAPI 3.1 document. The seed prints demo sign-in credentials (temporary passwords, shown once) for two isolated tenants, acme and globex — sign in at http://localhost:3000/login. Provision real tenants with pnpm tenant:create -- --name "Acme" --slug acme --admin-email admin@acme.com. Tear down (destroys volumes): pnpm dev:down. Full validation walkthrough: specs/001-platform-foundation/quickstart.md.

Demo data (for demos, screenshots, and videos)

To fill a tenant with realistic synthetic sample data (reports across states, entries, receipts, cards, approvals, budgets, analytics, a few fraud flags):

pnpm --filter @expense-platform/api seed:demo --tenant acme --volume default
#   --reset          clear the tenant's demo data first    --multi-entity   add a parent+children group
#   --allow-nonlocal (+ typed confirmation) required for any non-local database

It is a dev/demo tool only: it defaults to the local database and refuses a production-looking destination without an explicit opt-in, generates only fake data (no real PII/PANs/secrets), is idempotent, and tags its data as demo. Full options and safety behavior: specs/014-demo-data-loader/contracts/cli.md.

Project Status

Phase 1 (Platform Foundation) is implemented: multi-tenant accounts with row-level-security isolation, password + Google SSO sign-in, role-based access, the full 15-table expense schema with ISO reference data, audit logging, and a reproducible local environment. Phases 2+ (expense report lifecycle, receipts/OCR, card feeds, AI policy engine) are specified in initial-research/development-plan.md and follow incrementally. 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. The research notes flag GPL-3.0 (ERPNext) and LGPL-3.0 (Odoo Community) as relevant precedents in the open-source expense space; final licence selection will balance contributor reach against derivative-work obligations. See discussion for context.