The most analytically rich part of an annual report lives in the notes — unstructured, unqueryable, buried in PDFs. Equiinfra structures NTA data for Indian listed companies into a REST API, so analysts can query and compare across companies in milliseconds.
The most analytically rich disclosures — borrowing schedules, ageing buckets, contingent liabilities, related party flows — exist only as tables in 500-page PDFs. No API. No database. No query.
Five companies × five NTA types × three years = seventy-five tables read manually. Cross-company comparison at scale is economically impossible today for any analyst team.
Headline financials — revenue, EBITDA, EPS — are covered. NTA data is not. Borrowing instrument detail, ageing buckets, contingent exposure, RPT flows: a genuine whitespace in Indian market data.
We extract, canonicalise, and serve Notes to Accounts data via a REST API — so fund managers, analysts, and fintech builders can query NTA data the same way they query price data.
This data is live from our beta — pulled directly from the API. A fund manager gets this in milliseconds, not after reading three annual reports.
total_borrowingsshort_term_pct_of_totalecl_rate_pctdisputed_goodtax_disputes_totalperf_guaranteesnet_block_closingdividend_receivedsource_note, page_ref, and disclosure_variation flag — so analysts can verify any figure against the original annual report.
Each one addresses a real gap in how Indian listed company data is available today.
Exact as-disclosed extraction preserved in fs_raw. No normalisation that hides what the company actually reported. Source integrity is non-negotiable.
Beyond raw data — ECL rate trends with YoY basis-point deltas, maturity concentration signals, CWIP risk flags, governance signals on intragroup borrowings. Analyst-ready intelligence, not just numbers.
Canonical field names across all companies. One API call returns comparable data for every company in your query — no manual normalisation required. This is the core endpoint.
Every extracted figure carries a note number, page reference, and extraction confidence tier. Analysts can verify any number against the original annual report PDF — no black boxes.
Our founder is a Chartered Accountant. Extractions are built against actual Ind AS disclosure requirements — Schedule III, applicable accounting standards, sector-specific carve-outs. Not a generic PDF scraper.
When a company's disclosure deviates from the norm — a demerger, discontinued operations split, regulatory deferral account — it's flagged explicitly. No silent collapse of structurally different disclosures into a single misleading number.
Each NTA type is extracted into a canonical JSON schema — same field names across all companies, enabling direct comparison without manual normalisation.
Non-current and current borrowings by instrument. Maturity profile, security details, NCD outstanding, ICD governance flags.
Ageing buckets — 0–6 months to 3+ years. ECL rates, disputed amounts, allowance for credit loss movements.
Tax disputes, customs/excise claims, legal cases, performance guarantees on behalf of subsidiaries.
Transactions and balances with subsidiaries, associates, KMP. Dividend upstream dependency, ICD flows, management remuneration.
Gross block additions, disposals, depreciation. CWIP ageing — projects delayed over 5 years flagged. Asset age proxy computed.
Revenue, EBIT, assets by reportable segment. Geographic split. Inter-segment eliminations. Useful for conglomerates and multi-business companies.
Existing vendors cover XBRL-reported headline numbers — the minimum statutory disclosure. Notes to Accounts sit in PDFs, outside every existing data model. That is the gap Equiinfra fills.
Our founder is a Chartered Accountant. Every extraction is built against actual Ind AS disclosure requirements — Schedule III, applicable accounting standards, sector-specific carve-outs.
Structural variations are documented as canonical flags, not silently collapsed. L&T's subsidiary performance guarantee exposure, Infosys's zero conventional borrowings — each carries the right signal, not a generic number.
Each JSON carries a confidence_tier. Each structural departure carries a disclosure_variation flag explaining what's different and why.
Equiinfra is designed to sit alongside your existing data sources. A fund manager can use any terminal for price data and headline financials, and call Equiinfra for the notes layer — ageing analysis, debt structure, contingent exposure, RPT flows — feeding it directly into their own models. This is a non-overlapping data product with no current direct competitor in Indian markets.
Each company data package includes governance files — so any extracted figure can be verified against its source disclosure.
Maps every NTA type to its note number and page range in the original annual report PDF. Every extraction has a verifiable origin.
Structural fork flags for one-time deviations — demergers, regulatory deferral accounts, discontinued operations splits. Machine-readable and analyst-readable.
Confidence tier per field, variation flags for disclosures that deviate from the canonical norm, CA sign-off status on first-time extractions.
Building in phases, with market feedback driving priority at each step.
We're onboarding a small number of fund managers, analysts, and fintech builders to shape the product before we scale. Free API access during beta. Your feedback directly influences what we build next.