The financial rules engine. A new language for money.

International | 24K rows/sec | Self-hosted | EU AI Act Ready

Get Started
Early Access
Buy Licence

Deterministic

FinLang guarantees consistent, audit-grade outputs for every transaction. No randomness. No surprises.

Auditable

Every change is logged in JSON. Perfect traceability for teams and compliance workflows.

Fast

24,000 rows/sec on 5M×50 datasets. Performance that scales linearly.

View benchmarks

EU AI Act Ready

FinLang is engineered for transparency, explainability and auditability — core obligations of the EU AI Act (effective Aug 2026). Every decision is deterministic, logged, and reproducible.

Deterministic Logic

No black-box models — every rule is human-readable and predictable.

Full Audit Trail

Every classification produces a JSON diff for complete traceability.

Explainable Decisions

Outputs link directly to the exact rule that triggered them.

Learn about the EU AI Act →

Built for professionals. Trusted by developers.

FinLang uses a human-readable DSL for defining financial rules — clear, explicit, and fully auditable. It bridges simple, readable rules with enterprise-grade financial automation, bringing clarity to everything from personal finance to institutional reconciliation.

rule "Flag: Large Transaction (≥ 500)" {
  match:
  - amount in -1000000..-500
  - amount in 500..1000000
  set:
  - flags += "LargeTxn"
}

rule "Subs: SPOTIFY" {
  match:
  - counterparty ~ "*SPOTIFY*"
  set:
  - category = "Review"
  - flags += "Subscription"
}

Product Overview

A short walkthrough demonstrating FinLang’s engine, rules, CLI, and audit trail.

Under the hood: FinLang ships with a fully auditable CLI, rule engine, and feedback learning loop.

Deterministic execution

finlang --input in.csv --output out.csv --rules rules.fin

Consistent, audit-grade outputs for every run.

JSON audit trail

finlang ... --audit results.json --audit-mode lite

Every change logged for compliance and reviews.

Adaptive rule discovery

finlang-discover --input categorized.csv --candidates out.csv

Learn from data to propose new rules safely.

Early Access Programme

To help shape the future of FinLang, we’re inviting a small number of organisations to join our Early Access programme where teams can help shape the FinLang DSL and contribute rulepack patterns.