Review Your Flow of Funds With Formance
Walk through how a programmable core ledger fits your specific flow of funds, deployment model, and compliance scope.
Modern Treasury built its reputation on US bank connectivity, automating ACH, wires, and reconciliation for teams moving fiat between US bank accounts. That focus is a strength for US-only treasury operations, but it is a ceiling for teams whose money problem is bigger than a US payment-operations platform.
Five triggers push teams to evaluate Modern Treasury alternatives, and most reach out after hitting two or three at once:
The five best Modern Treasury alternatives in 2026 are Formance, TigerBeetle, Fragment, Stripe's ledger primitives, and building in-house, benchmarked against Modern Treasury.
| Property | Modern Treasury | Formance | TigerBeetle | Fragment | Stripe's ledger primitives | Build in-house |
| Layer in the stack | Payment operations plus managed ledger | Core ledger plus Connectivity, Flows, Reconciliation | Specialized OLTP storage engine | Hosted core ledger API | Ledger-like records inside Stripe | Whatever you build |
| Open source | Proprietary, vendor-managed | MIT (Ledger, Connectivity) | Apache 2.0 | Proprietary, vendor-managed | Proprietary, vendor-managed | Your code |
| Multi-asset | Fiat plus stablecoin rails via MT Payments | Native fiat and digital assets in one ledger | Asset-agnostic via ledger/account partitioning | Model-defined, hosted only | Stripe-denominated balances only | If you build it |
| Self-hosted | No, hosted-first | Yes, on your own cloud | Yes, self-managed cluster | No, hosted-first | No | Yes |
| Reconciliation depth | Deep bank-side recon, many-to-many rules | Automated matching, drift detection, settlement error alerts | External layer required | Application-layer work required | Stripe-scope only | Custom scripts you maintain |
| Compliance certifications | SOC 2, bank-vendor posture | SOC 2 Type II, ISO 27001, DORA compliant, plus support for MiCA and GENIUS frameworks | Assess against your requirements | Assess against your requirements | No separate ledger product to certify | Your audit to pass |
| Pricing model | Usage-based enterprise pricing, contact sales | Open-source core (free) plus enterprise tier | Open-source (free), self-hosted operating cost only | Hosted tiered pricing, contact sales | Included with Stripe products | Internal engineering cost |
| Time to production | Weeks to months depending on bank onboarding | 2 to 4 weeks per custom connector, days for pre-built | Months to build a ledger product on top | Under 30 minutes to first simulation | Immediate for Stripe-native stacks | 6 to 18 months to audit-ready |
| Geographic rail coverage | US-dominant, limited non-US bank partners | Global via PSP, bank, exchange, and custodian connectors | Rail-agnostic, you integrate rails yourself | Rail-agnostic, you integrate rails yourself | Stripe's supported countries only | Whatever you integrate |
Formance is an open-source, programmable core ledger platform that unifies fiat and digital assets with regulatory-grade traceability. The ledger offers a real-time double-entry engine with bi-temporal history and append-only postings.
It also includes modules for Connectivity, Flows and Reconciliation.
Formance is a system of record and not a money movement platform. Formance does not hold or move funds itself. All money movement flows through connected payment providers (PSPs, banks, exchanges, or custodians).
The distinguishing primitive is Numscript, a purpose-built DSL for describing multi-party financial transactions. You write the transaction once, and Numscript handles the fee splits, sub-account movements, and currency conversions as a single atomic operation. Every leg commits together, or none commits at all, so the ledger never lands in a partial state.
A marketplace payout that splits a $100,000 customer payment across a merchant, a platform fee, and a tax-holding account looks like this in Numscript:
send [USD/2 10000000] (
source = @world
destination = {
80% to @merchant:acme
15% to @platform:fees
5% to @tax:holding
}
)
The three destinations settle atomically. If any leg fails, none of the postings commit, so the ledger cannot end up in a partial state where the merchant is credited, but the tax holding is not.
Formance's connectors directly counter Modern Treasury's US-centric rail depth and Fragment's zero-connector model. Pre-built connectors span eight categories (PSPs, cross-border, BaaS, direct-API banks, Open Banking, digital asset custody, crypto exchanges, and direct on-chain rails), covering vendors like Stripe, Wise, JP Morgan, Fireblocks, and Coinbase.
For providers outside the pre-built list, teams typically build a custom integration in 2 to 4 weeks using the open-source generic connector framework.
TigerBeetle is a purpose-built OLTP database for financial transactions, written in Zig with a fixed schema tuned for accounting workloads. It is a storage engine and not a direct ledger platform.
Treating TigerBeetle as a drop-in replacement for Modern Treasury Ledgers or Formance Ledger will fail. Treated as a fast, correctness-focused backing store for a ledger you build on top of, TigerBeetle earns its place.
The engine is built for multi-object transactions under high contention. TigerBeetle co-locates the accounting logic with the storage and runs as a six-replica distributed system, reducing each transfer to a handful of in-memory operations. The result is what the docs call the "speed-of-light" path for financial state changes.
TigerBeetle is designed to run alongside a general-purpose database like Postgres. TigerBeetle handles accounts, transfers, and balance updates in the hot path. Postgres (or Formance, or your app database) stores everything else: metadata, user records, and business logic.
Fragment is a hosted ledger API positioned as the fastest path from "we need a ledger" to "the ledger is running." Fragment's differentiator is a visual ledger simulator ('Scenarios') that lets teams model a chart of accounts and simulate transaction flows interactively before writing code. Fragment's own materials claim teams can set up the full ledger infrastructure in under a week.
It is useful during evaluation and replacing a typical two-week proof of concept (PoC). A short evaluation build where a team validates whether a vendor's product actually works for their use case before committing to a full integration. Stripe backed Fragment's $9 million seed in July 2024, and the same report places total funding at $10.8 million.
Fragment's product surface is a ledger DSL, SDKs in multiple languages, sandbox and production environments, and a hosted infrastructure model. Fragment is the shortest path to production for early-stage teams that need a real double-entry ledger without running infrastructure or waiting through a six-month platform build.
Fragment's limits show at integration boundaries. Fragment lacks Modern Treasury's provider connectivity and Formance's multi-asset breadth, and complex multi-rail orchestration lands in your application layer. Fragment's team, funding, and enterprise readiness are also earlier-stage than mature vendors.
Stripe's ledger primitives (balances, transactions, Connect account records, and Issuing authorizations) are tied to Stripe's payment products and live inside Stripe's platform. Stripe has extended this with Financial Accounts for Platforms (formerly Stripe Treasury), which adds FDIC pass-through insured accounts, direct ACH and wire transfers, multi-currency balances, and stablecoin-backed balances in 100+ countries.
But even with that expansion, Stripe's primitives remain a record-keeping layer for money that moves through Stripe's own network of bank and rail partnerships. They are not a general-purpose ledger.
The moment money moves through a rail or partner Stripe doesn't support natively (a direct bank relationship outside Stripe's partner network, a non-Stripe merchant of record, RTP or FedNow access Stripe doesn't offer, or a PSP) Stripe's primitives stop being sufficient as a system of record.
You now have "Stripe balances" and "not-Stripe balances," with no atomic ledger across both, and discovering the need for a unified ledger at that point is costly.
Building your own ledger means owning every layer (concurrency, idempotency, multi-party transactions, bi-temporal history, reconciliation against external providers, and regulatory-grade traceability) as an ongoing engineering commitment rather than a one-time build. Any competent team can build a ledger that works in a demo. Keeping the ledger working through years of scaling events, new rails, and audit cycles is the harder problem.
The honest question isn't "can we build a ledger?" It's "is the ledger the product we sell?" If the ledger is the product, build it. If not, every engineer-week spent on ledger internals is one not spent on the product customers pay for.
Audit-readiness is a separate milestone from "works in production." SOC 2 controls, evidence collection, immutability proofs, and a documented reconciliation process are what turn a working ledger into an auditable one.
Choose the alternative that matches where your money problem lives: in a programmable multi-asset core (Formance), in a high-throughput engine (TigerBeetle), in a fast-start hosted ledger (Fragment), inside Stripe (Stripe primitives), or in a custom build.