Line 02 · Fintech · Opdrik Recon

N-way reconciliation, built config-not-code.

An enterprise reconciliation platform for banks, fintechs and payment institutions — transaction matching, automated fee & tax assurance, governed break management, and regulator-ready reporting.

CORE SWITCH SCHEME 3-WAY MATCH ✓
The idea

Stand up a new reconciliation rail without an IT project.

Opdrik Recon is built on a declarative, config-not-code design. New workflows are defined in a functional UI with strict four-eyes publication — no long-cycle development. Every configuration compiles into efficient, set-based SQL.

Config, not code

A new "rail" is stood up through the UI with maker–checker publication, eliminating bespoke development for each workflow.

Dual-engine, identical results

The same declarative config runs natively on both PostgreSQL and Microsoft SQL Server, verified for byte-identical results by a cross-engine test suite.

Runs inside the database

Matching executes as set-based SQL in the data layer — incremental, so already-matched rows are never rescanned.

Verified benchmarks

Measured, not projected.

Figures below come from large-scale synthetic test runs with hard, embedded ground truth. We publish empirical numbers rather than theoretical ceilings.

16×
faster heavy 3-way runs — 5,495s reduced to 334s, staying flat as data accumulates.
~36s
typical multi-rail day-run execution time.
20
foundational zero-code rails supported out of the box.
7–10s
per-day bulk ingestion, validated to 1,570 files and ~1.85M rows in a synthetic test month.
3dp
fee and tax matching precision — invoices reconciled to three decimal places.
1
immutable Attestation Pack PDF generated per reconciliation run.
Capabilities

Matching, assurance, and governed break management.

Zero-code rails, sourced

Out-of-the-box support for 20 foundational rails: General Ledger and suspense accounts, card networks (3-way core-banking × switch × scheme), instant payments, ACH, RTGS, cheque clearing, and biller networks.

Incremental matching

Rows are never rescanned once matched — processing cost is restricted to new arrivals plus residual breaks, so run times stay flat as monthly data grows.

Fee & tax assurance

Fee schedules and tax matrices are versioned, effective-dated data. The engine recomputes every expected fee rather than trusting provider statements, matching monthly invoices to three decimal places.

India compliance — GSTR-2B / ITC

An automated Input Tax Credit recovery mechanism: monthly supplier-invoice audits run against effective-dated GSTR-2B datasets to isolate mismatches and leakages.

Governed break management

Discrepancies become persistent, uniquely tracked cases carrying exact monetary exposure. Aging buckets, priority queues (overdue → age → exposure) and per-band SLA alerts are fully configurable — including RBI Turnaround-Time penalty clocks (e.g. ₹100/day).

Attestation & audit

An append-only, hash-chained audit trail. Every run auto-generates a single immutable Attestation Pack PDF with the active rule version, maker≠checker approvals, break evidence, and the full audit log.

How a run works

Ingest, compile, isolate, attest.

01 — INGEST

Ingest & stage

External pipelines (e.g. Apache Hop or your existing corporate ETL) feed a strict staging contract — zero custom client code inside the core engine.

02 — COMPILE

Compile & run

Declarative rules compile into set-based SQL, executing high-speed incremental N-way matching inside the database layer.

03 — ISOLATE

Isolate & escalate

Unmatched exceptions instantiate persistent cases, assigned to prioritised queues by aging threshold and financial risk band.

04 — ATTEST

Approve & attest

Actioned items pass through configurable maker-checker gates, compiling into a tamper-evident, auditor-friendly Attestation Pack.

On the roadmap · UPI high-volume scale

A specialised UPI module, engineered and in load-testing.

An imminent, concrete architectural extension — designed to reconcile high-volume mobile payments as the governed tier of record beneath wire-speed auto-match.

Engineered to process tens of millions of records per day.
Advanced 3-way and 4-way matching matrices for distributed mobile transactions.
The governed tier into which streaming residues drain — the single source of policy and audit truth, making automated write-offs defensible to examiners.
Where this stands, precisely. The core engine is fully built and rigorously validated on large-scale synthetic datasets with embedded ground truth. Full production-volume certification on SQL-Server-primary deployments and the ultra-high UPI-scale sustained bands (~5M+ transactions/day) are in active, scheduled load-testing. We publish verified benchmarks, never theoretical projections — so this is presented as an imminent extension, not a live production claim.
Deployment & governance

Runs where your data lives.

InfrastructureOn-premises or secure private cloud
DatabasesPostgreSQL · Microsoft SQL Server
Client-specific codeNone — external staging + UI-driven config
Audit trailAppend-only, hash-chained
For evaluation teams

Request the full performance evidence pack.

Technical evaluation groups and architecture panels can request access to complete benchmark and attestation evidence.