We built a private, internally-hosted platform for a client who manages large volumes of GST sales and purchase data. What used to be days of spreadsheet wrangling is now a single upload — the system cleans, validates and files every record, then turns years of transactions into reports on demand.
The problem: reconciling GST data by hand
The client's data arrived the way real business data always does — messy. Every vendor sent files in a different shape: modern Excel workbooks, legacy spreadsheets, CSVs, plain text, even PDFs. Someone had to open each one, copy rows into a master sheet, fix the formatting, and hope nothing was missed.
That manual process created a stack of recurring pain:
- Re-typing and copy-paste errors across large, multi-vendor datasets.
- Inconsistent file formats that each needed handling by hand.
- Duplicate and variant vendor and party names — the same business entered three slightly different ways — quietly fragmenting the data.
- Dirty numbers (currency symbols, stray commas) that broke totals.
- Silent failures — when a row didn't import, nobody knew which one or why.
- Slow financial-year roll-ups that meant rebuilding the same report every time someone asked.
What we built
A single pipeline that takes a raw file and turns it into clean, queryable data.
- One import for every format. Excel, legacy spreadsheets, CSV, text and PDF all flow through the same bulk-import pipeline, so the team stops handling each vendor differently.
- Automatic de-duplication. Vendors and parties are matched on their GST identity and normalized names, so "MAHARASHTRA SEAMLESS LIMITED" and "Maharashtra Seamless Limited" collapse into one record instead of two.
- Row-level validation you can trust. Every row is checked against GST rules. Anything that fails is flagged with a plain-English reason and written to a downloadable "rejected rows" file — so the team can fix just those rows and re-upload, instead of guessing what went wrong.
- Imports that don't freeze the screen. Large files are processed in the background with live progress, and every import is logged with its running totals in an import-history dashboard.
- A searchable data browser. Filter by vendor, party, GST number, month, financial year or custom period — no more scrolling a spreadsheet.
- Reports and exports on demand. Report totals and a financial-year pivot (parties against financial years) are generated on request and exported to CSV — the roll-up that used to be rebuilt by hand is now a click.
Built to be trusted with financial data
Because this platform holds sensitive financial records, it's engineered for control and durability, not just features:
- Role-based access — administrators manage data and imports; viewers get safe, read-only access.
- Strong authentication — two-factor login and passkeys (WebAuthn), not just a password.
- Resilient processing — background jobs with progress tracking, and failures caught and surfaced instead of lost.
- Daily automated backups of the database, so the data is always recoverable.
The outcome
The manual reconciliation loop is gone. Files go in, clean and validated records come out, and reports are ready whenever they're needed — with a clear, fixable trail for anything that doesn't fit. The team spends its time on the numbers that matter, not on re-typing them.
Want something like this?
We build custom software and automation that removes the manual glue around your data — at a fixed price, starting with one clear win. Tell us the workflow you dread most, or see how we price projects.