Can Reconova compare two Excel or CSV files?

The browser workspace profiles and cleans CSV, TSV, and JSON files locally. Secure cloud processing profiles CSV, TSV, JSON, NDJSON, XLSX, and Parquet today; exact and fuzzy two-source reconciliation arrives in Phase 6.

Does Reconova upload my CSV data?

Not in browser mode. Local profiling and exports stay on your device. If you explicitly choose Secure upload while signed in, the browser sends the file directly to private object storage using a short-lived permission bound to one random object key.

How does the data-quality score work?

The score combines completeness, validity, consistency, uniqueness, and conformity. Reconova shows each factor separately so you can see why the score changed instead of relying on an unexplained number.

Will Reconova preserve IDs with leading zeros?

Yes. Values such as 001234 and long account references are treated as identifiers by default. A destructive type conversion must be added explicitly, previewed, and configured with an error behaviour before export.

Which formats work in the browser today?

The local workspace supports CSV, TSV, and JSON files up to 25 MB and 100,000 rows, subject to browser memory. Secure server profiling supports CSV, TSV, JSON, NDJSON, XLSX, and Parquet up to 100 MB.

Motifuse workspace

Make business data fit for decisions.

Reconova is a professional data-quality, transformation, and reconciliation studio. Understand broken values, build a reviewable cleaning pipeline, and preview the impact before anything leaves your browser.

Private browser processing
Non-destructive dry runs
Identifier-safe inference
Virtualized data grid

Reconova Studio

Browser workspace + secure cloud profiling

Files stay on this device Undo & redo

Private browser workspace

Turn messy business data into a trusted dataset

Upload a CSV, TSV, or JSON file. Reconova profiles every column off the main thread, explains quality issues, and lets you dry-run safe cleaning steps before export.

Inspect quality

Nulls, uniqueness, types, duplicates, and likely PII.

Build a pipeline

Ordered, reversible steps with strict validation.

Preview and export

Measure the impact before creating a safe result.

Browser mode keeps file bytes on your device. Secure upload is opt-in, direct to private object storage, and clearly separated from the local workspace.

Explainable quality

A score you can actually investigate

Reconova does not hide data quality behind a decorative number. Every score is broken into factors, and every column exposes the evidence that affects it.

Completeness

How much expected data is present rather than blank or null.

Validity

Whether rows parse cleanly and values follow safe structural expectations.

Consistency

How strongly values agree with the detected type and format.

Uniqueness

Exact duplicate pressure across the dataset and potential keys.

Conformity

Whitespace, Unicode, identifier, and column-pattern findings.

Real work, not a converter

A disciplined workspace for messy operational data

Reconova is designed around the questions analysts, finance teams, operations teams, consultants, and small businesses ask before data reaches another system.

Clean a customer CSV

Find missing email addresses, inconsistent status values, extra whitespace, duplicate records, and identifier risks before a CRM import.

Prepare finance data

Normalize vendor names, inspect amounts and dates, preserve invoice references, and send rejected conversions to a reviewed exception path.

Validate a migration

Profile an export before import, compare expected types, locate null-heavy fields, and create a repeatable transformation sequence.

Find duplicate records

Measure exact duplicate rows now, then extend the same project into composite and fuzzy matching in the reconciliation phase.

Reconciliation architecture

Compare datasets without an O(n²) guessing game

The server reconciliation phase will normalize values, match exact and composite keys, block sensible candidate groups, score fuzzy fields, and route uncertain matches to human review. Every result will include a reason—not just a confidence badge.

Exact and composite keys
Date and amount tolerances
Candidate blocking
Weighted fuzzy fields
Ambiguous-match review
Unmatched and exception exports

Illustrative match explanationReview needed · 86%

Vendor ID exact matchsupport

Vendor name 92% similarsupport

Invoice amount within ₹10support

Invoice date within two dayssupport

Email domain differsconflict

This review flow is intentionally not active in Phase 1. It will arrive after the storage, worker, profiling, and transformation phases are stable.

Formats and limits

Use the right processing mode for the file

Small text datasets benefit from instant local feedback. Complex workbooks, columnar formats, large exports, and reconciliation jobs need controlled server processing with durable progress and cleanup.

Privacy by processing mode

Browser mode makes no file-content request. Secure mode uses private Object Storage, exact-object short-lived upload permissions, tenant-scoped random keys, explicit deletion, and enforced retention limits.

Format	Browser mode	Server phase
CSV	Profile, transform, export	Full profile + canonical Parquet
TSV	Profile, transform, export	Full profile + canonical Parquet
JSON	Array-of-objects datasets	Full profile + canonical Parquet
XLSX	Not in browser	Sheets, masking, canonical Parquet
NDJSON	Not in browser	Invalid-row report + profile
Parquet	Not in browser	Native profile + canonical Parquet

Frequently asked questions

Know the boundary before you process

Reconova is explicit about what runs today, what remains local, and what will require secure server infrastructure later.

Keep working

Related Motifuse tools

Use these focused utilities when you need a quick one-off conversion, comparison, or text cleanup outside a Reconova project.