Template-driven two-tier matching between partner CSV files and your ledger. Configurable column mappings, timezone tolerances, regex and JSONPath extraction, MD5 duplicate detection, and partition-pruned joins across billions of rows. Strict matches (all criteria met) plus a loose fallback for identifier-only agreement.

Import partner CSV files and automatically match them against your ledger using configurable templates. Manages exceptions, multi-account matching, and currency validation — built for Reconciliation Officers handling high-volume settlement processing.

Automate recurring financial operations using pg_timetable. Schedule account snapshots, partition maintenance, reconciliation sweeps, and balance roll-ups — all configurable and monitored from a single administrative view.

Define inheritable configuration templates applied across categories of accounts. Each profile encodes 100+ parameters — velocity limits, balance thresholds, transaction size controls, and fraud flags — enforced automatically at transaction time.

Build and manage your chart of accounts with hierarchical parent/child aggregation, multi-currency support, and timezone-aware reporting dates. Supports Asset, Liability, Income, and Expense account types with automatic real-time balance propagation.

Proprietary cyclic fraud detection using trigonometric behavioural analysis. Profiles 90-day statistical baselines per account across hour, day-of-week, and day-of-month dimensions to surface statistical outliers in real time.

Trigger-based audit log capturing every insert, update, and delete across all system tables. Each entry records the user, role, client IP, action, and a full JSON diff of before/after values — stored in security.system_audit.

Integrate with any OIDC-compliant identity provider — Keycloak, Auth0, Microsoft Entra, or your own IdP. Federated identities are mapped to internal roles, preserving the platform's granular 25+ permission model.

Create and manage local platform users and role assignments independently of, or alongside, your SSO configuration. Every user is governed by 25+ permission categories with r / rw / rwd granularity.

Generate paginated, filterable statements showing all postings, running balances, and transaction references for any account. Supports Accounting Managers, Customer Service Representatives, and end-user self-service.

Analyse account fund movements, velocity trends, and throughput patterns over configurable time windows. Visualises inflow and outflow dynamics for treasury monitoring and operational reporting.

Periodically commits a SHA-256 root hash of all account balance chains to the Bitcoin blockchain via OpenTimestamps. The deterministic CSV snapshot is the source of truth — independently verify any anchor with nothing more than sha256sum and the OTS CLI.

Each confirmed anchor produces a complete .ots proof file containing the Merkle path from your root hash to a specific Bitcoin block. Download the proof, inspect the Bitcoin block height, and verify the OP_RETURN commitment independently of this system.

A mobile-first customer wallet (installable PWA + Android TWA, mirrored in SQLPage) on top of the same double-entry ledger. Six money movements — Send (P2P or B2B merchant), Request, Top-up, Settle, and Pay-by-QR — through one self-scoped api.me_* access layer. A configuration-driven fee engine prices every operation independently (threshold, percentage/absolute, cap, payer/payee/both, deduct-on-top vs deduct-from), posting each fee as its own auditable ledger leg. Runs under the customer's own low-privilege role with row-level security, ownership-gated definer RPCs, and a strict-CSP front-end.

Automated onboarding compliance across five verification types — KYC, KYB, KYM, KYT, and CDD. Country-specific requirement sets, three verification tiers, Docling OCR document processing, and a fully automated invitation-to-approval workflow with 7-day reminder cycles. Staff review dashboard and customer self-service portal included.

Conversational AI embedded in the platform sidebar. Multi-turn sessions with rolling context, retrieval-augmented generation against platform documentation, and MCP (Model Context Protocol) tool use for live ledger queries — all governed by your existing row-level security policies. Ask Flo from any page with a single click.

Eight-dimension behavioural profiling using trigonometric circular statistics. Builds 30-day and 90-day per-account baselines across time-of-day, day-of-week, day-of-month, transaction amount, and velocity dimensions. Composite score (0.0 – 1.0) flags accounts the moment a transaction deviates statistically — no rules to write, no ML infrastructure to deploy. Entirely inside PostgreSQL.

Monitor transaction write latency and execution time distributions across the ledger engine. Identifies performance bottlenecks and validates system responsiveness under load — essential for capacity planning.

Track transaction counts, volumes, and trends over time. Provides operational visibility into peak periods, processing capacity utilisation, and growth trajectories.

A complete OpenTelemetry + LGTM stack (Grafana, Loki, Tempo, Mimir/Prometheus, Pyroscope) ships with the platform and starts with it — traces, metrics, logs, and continuous profiling correlated in a single Grafana pane. Follow one request edge → REST API → PostgreSQL, then jump from a latency spike to the exact trace, its logs, and the CPU flame graph of the responsible function. No external SaaS, no extra setup.

A self-hosted Wazuh + OpenSearch SIEM watches the platform that handles the money. Four source classes — container/host security, the tamper-evident database audit trail, edge web-attack detection, and file-integrity & vulnerability monitoring — feed one detection engine. Severity-tiered alerts are mapped to MITRE ATT&CK and tagged for PCI-DSS, GDPR, HIPAA and NIST, so detections double as continuous compliance evidence.

Flo (Florence) is the conversational AI embedded directly in the SQLPage admin UI — a persistent sidebar on every page, ask-from-anywhere. She grounds every answer two ways. First, retrieval-augmented generation (RAG): before responding she searches the embedded platform-documentation corpus by vector similarity and injects the most relevant chunks, so answers reflect how this deployment actually works. Second, live Model Context Protocol (MCP) tool use across three tool families — LedgerFlow (accounts, transactions, balances, reconciliation, fraud scores), Grafana (observability dashboards, metrics, traces, logs) and Wazuh (SIEM alerts and security posture). Every call runs under your own token and the platform's row-level security, so Flo never sees write/admin surface or data you cannot. The Grafana and Wazuh tool families are permission-gated (j_observability_r) and default off.

The fastest way for customers to integrate their own platform into LedgerFlow. Point the connector feature of your existing tool — VS Code (Copilot Agent), Claude Code, Claude Desktop, Cursor — at this deployment over the public Model Context Protocol server. A single read-only token lets the agent explore the live schema and documentation and write correct integration code, turning days of doc-reading into an afternoon. Curated and RLS-bounded: the agent never sees write or admin surface. Generate a token and copy ready-made client config on the Setup → Agent & MCP page.