I'm reconciling historical NSE:INFY EQ (token `408065`, ISIN `INE009A01021`) data for 11 November 2026. The read-only Kite historical API produced 360 minute candles for **09:15–15:14 IST**, with total minute volume **8,082,671 shares**. The official-format NSE daily bhavcopy for that session reports **9,148,933 shares**.
A separately sourced 1-minute feed supplies all 360 matching timestamps but reports **8,294,256 shares** over the same 09:15–15:14 window (difference **211,585 shares**, approximately 2.62% of Kite's total). Of 360 aligned candles, **349** differ in volume, and **151** differ in close, with mean absolute close difference ₹0.0786 and max ₹0.80. Fixed 0-minute alignment is materially closer than ±1/±2-minute offsets.
I'm aware of the earlier Kite forum explanation that CAS-eligible securities stop regular minute bars at 15:14, and that daily bhavcopy and historical intraday bars come from different sources. That explains some daily-vs-intraday variation; I would like to understand whether it also explains the above **same-window, two-provider** difference.
Could you confirm: 1. Are Kite minute OHLCV candles constructed from a sampled tick feed, per-trade messages or corrected exchange transaction data? Are late trade/candle revisions applied? 2. Are minute bars timestamped at the start of each minute, in IST, and is the 15:14 bar the last ordinary market bar for INFY on this session? 3. Does Kite exclude particular exchange trades, odd-lot/cross-market activity or CAS-related prints before 15:15? Is there a documented expected reconciliation gap? 4. Do historical minute EQ candles receive dividend adjustments, including the November 3 INFY ₹26 ex-date? Are intraday prices adjusted retroactively? 5. Is there an official candle-generation specification or market-data source attestation we can cite in an audit?
I'm seeking metadata and methodology, not trading signals. No API secrets, tokens or account details are included.
