GST invoice to JSON for software and automation

Software does not read invoices; it reads fields. Your accounting system, internal ERP, or custom workflow wants a GSTIN, an invoice number, a taxable value, and a tax split as keys and values, not as text on a page. Dynamite Docs extracts GST invoices, PDFs, scans, or photos, into clean JSON with the schema inferred per layout and consistent across a batch. A REST API with tokens lets your systems send documents and pull structured rows, and a webhook fires when each extraction completes, so the pipeline runs without a human exporting a file.

Try this workflow free or view pricing.

The invoice is a document; your system needs data

Accounting software and internal systems consume JSON, not spreadsheets or PDFs. The invoice has to become structured keys, and the field names have to be consistent across suppliers for any of it to be useful.
A GST invoice carries a fixed field set, GSTIN, invoice number and date, vendor and customer, HSN/SAC, item lines, taxable value, CGST, SGST, IGST, and totals, but the layout differs per vendor, so a template approach breaks on the first new format.
Documents arrive in mixed quality: clean digital PDFs that need no vision, and scans and photos that do. A pipeline needs to route each one without someone deciding file by file.
Automation only helps if it triggers on its own. If someone still exports a file and mails it to the system, the bottleneck moved but did not disappear.

How GST invoices become JSON your systems can consume

01 — Send the documents in

Upload invoices in the workspace or send them through the REST API with your API token. Folders, single documents, and batches of hundreds are handled the same way.

02 — Schema inferred and consistent

The GSTIN, invoice number, date, vendor and customer, HSN/SAC, item lines, taxable value, the CGST/SGST/IGST split, and totals come out with the same field names across every layout, inferred on the spot, no templates.

03 — Route deterministic or vision

Clean digital PDFs parse deterministically from their text layer at no cost; scans and photos are read by the vision model you choose. Each document takes the cheaper path that works.

04 — Webhooks carry the result

When an extraction completes, a signed webhook posts the structured rows to your endpoint, or your system pulls them from the API. The schema stays stable, so downstream automation does not need to change per vendor.

Document types this workflow handles

GST invoice PDFs — digital invoices parsed deterministically
GST invoice photos — vision OCR for scans and phone photos
Bulk Tally and billing-software exports — text-layer PDFs, consistent output
Credit / debit notes — adjustments as structured rows

What comes out as JSON keys

The same key names across the batch, so downstream code does not change when the vendor does:

GSTIN (supplier) — 29AABCC3456C1ZD
Invoice number — KT-5512
Invoice date — 2026-08-07
Vendor / customer — Ganesh Industries
Place of supply — Karnataka (29)
HSN / SAC code — 8467
Item & quantity — Rotary hammer × 6
Taxable value — ₹1,86,000.00
IGST (rate · amount) — 18% · ₹33,480.00
Invoice total (incl. tax) — ₹2,19,480.00

A realistic example

The workspace shows the same data the API returns. A batch extracts to rows like these, each one also available as JSON with stable keys:

Invoice #	Date	Vendor	GSTIN	HSN/SAC	Item	Taxable value	CGST	SGST	IGST	Total
KT-5512	2026-08-07	Ganesh Industries	29AABCC3456C1ZD	8467	Rotary hammer	₹1,86,000	—	—	₹33,480	₹2,19,480
INV/2026-27/0188	2026-08-11	Mehta Steel Supplies	27AABCT1429P1ZM	7210	GI sheets	₹5,40,000	₹48,600	₹48,600	—	₹6,37,200

Stable schema, automatic routing, no retyping

The value of JSON over a spreadsheet is that it feeds systems directly. The schema is inferred per layout, but the output is consistent across a batch: the same key names for GSTIN, invoice number, invoice date, vendor, customer, HSN/SAC, item lines, taxable value, CGST, SGST, IGST, and total, whether the source was a Tally export, a billing-software PDF, or a photographed invoice. A stable schema means your accounting software, internal ERP, or automation does not change when a vendor changes their layout, which is the failure mode a template library cannot escape.

Routing is automatic because it is decided by the document, not by a person. A clean digital PDF carries a text layer and parses deterministically, identical results every time, at no cost. A scan or photo without a text layer goes to the vision model you choose. The same endpoint handles both, and the cost of the batch tracks the documents in it.

Webhooks are what turn extraction into automation. Each completed extraction posts a signed payload, an extraction.completed event, to your endpoint, so an invoice entering the system can trigger posting, matching, or any downstream step without a file being exported and forwarded. The rows stay reviewable in the workspace for the human check that matters, and when you prepare GSTR-1 you work from data you can trace to the source document. Consult the GST portal and your tax advisor for current rules.

Why bring-your-own-AI matters for automation

Automation only works if the schema is stable. GSTIN, invoice number, HSN/SAC, taxable value, CGST, SGST, IGST, and totals come back under the same key names from every vendor, so the downstream code never changes when a layout does, and the model that reads each page is your choice, at your provider’s rate with zero markup.

The routing is automatic and cost-aware: clean digital invoices parse deterministically at no cost, scans and photos go to the vision model you pick, and a signed webhook posts the completed rows to your endpoint without a human exporting anything.

The schema, the webhooks, and the corrections your team makes are the pipeline asset, not a vendor’s product. Export the same rows as JSON, CSV, or Excel whenever a system prefers a file, and keep the trace back to the source document for GSTR-1.

Questions, answered

Is there a REST API for sending invoices and getting JSON?

Yes. The REST API accepts documents with your API token and returns structured rows. The same extraction is available in the workspace for review, and exports cover Excel, CSV, and JSON.

Do webhooks fire when extraction completes?

Yes. Each completed extraction can post a signed payload, an extraction.completed event, to your endpoint, so downstream systems and automation trigger without a manual export.

Is the JSON schema consistent across vendors and document types?

Yes. The schema is inferred per layout but kept consistent across a batch, so GSTIN, invoice number, HSN/SAC, taxable value, CGST, SGST, IGST, and totals use the same field names from every source.

Does it route deterministic and vision documents automatically?

Yes. Clean digital PDFs parse deterministically from their text layer at no cost; scans and photos are read by the vision model you choose. Both paths return the same schema through the same API.