Deterministic Reconciliation. AI-Assisted Review.
Settler is API-first infrastructure for finance and fintech teams: a deterministic matching core, AI-assisted exception triage, tenant-safe isolation, and audit-grade evidence trails built into every run.
Data In → Deterministic Engine → AI-Assisted Review → Verifiable Audit Trail
Platform Architecture
Deterministic Engine
Rules-based matching with field-level tolerance. Same inputs and same rules always produce identical output and identical evidence hash. No hidden logic.
- Configurable matching rules in code
- Field-level variance with tolerance windows
- Multi-source joins across adapters
- Floating-point safe (integer cent arithmetic)
AI-Assisted Review
AI compresses exception triage time by surfacing context, suggesting classifications, and flagging anomalies. Humans retain final authority on every decision.
- Exception context summarization
- Anomaly pattern detection
- Suggested resolution classifications
- Review history and audit trail
Governance & Controls
Role-based access, tenant isolation, and configurable policy controls. Every decision is logged with actor, timestamp, and documented reason.
- Owner / Admin / Reviewer / Viewer roles
- Tenant-scoped API keys
- Policy-gated mismatch resolution
- Exportable compliance evidence
Replay & Provenance
Any historical run can be replayed from stored inputs and rules. The output hash must match the original. Debug mismatches or prepare for audit without re-running live systems.
- Hash-verified replay of any run
- Rule version fingerprints in evidence
- Provenance chain across sequential runs
- Determinism CI enforcement
Integrations & Adapters
Connect the systems you already use. Adapters handle schema normalization so your rules stay clean.
Self-Host or Managed
Deploy Settler inside your own infrastructure. Your financial data never transits a Settler-managed network unless you choose the managed cloud deployment.