indie market analyst
an open-source, local-first indian-market analysis toolkit — scanning, dashboards, backtesting, and optional ai chat.
what it is
a full-stack market workspace for nse/bse, meant to run on your own machine. the ai layer is genuinely useful, but it's one feature of the toolkit — not the whole product. the deterministic parts work, and are testable, without any llm key at all.
four surfaces
the parts i'm proud of
the intraday engine fetches candles, applies liquidity and asm/gsm filters, runs orb, vwap-reclaim, breakout, flag, momentum, and short-cover detectors, then scores qualified picks. the backtester is a pure pandas/numpy engine with the real indian cost model — stt, stamp duty, exchange transaction charges, sebi fees, gst, and zerodha-style brokerage — because a backtest that ignores costs is fiction. nse direct is the primary data source, with a yfinance fallback.
a trade ticket, with every charge
enter one round trip on nse and see what it really costs. the default is a "profit" of 5 rupees that the charges turn into a loss.
with javascript off, here is the default trade worked by hand: buy 10 at ₹500, sell at ₹500.50, intraday. gross ₹5.00. brokerage ₹3.00, stt ₹1.25, exchange charges ₹0.31, sebi ₹0.01, stamp duty ₹0.15, gst ₹0.60. charges ₹5.32. net −₹0.32.
rates are zerodha's retail schedule as i read it off their charges page on 10 august 2026: brokerage 0.03% or ₹20 per order, whichever is lower (zero for delivery); stt 0.025% on the sell side intraday, 0.1% both sides delivery; nse transaction charges 0.00307%; sebi ₹10 per crore; stamp duty 0.003% (0.015% delivery) on the buy side; gst 18% on brokerage, exchange and sebi charges. delivery dp charges are left out. not advice; check your broker's current schedule.
built by dharun ashokkumar