Indie Market Analyst
An open-source, local-first Indian-market analysis toolkit — scanning, dashboards, backtesting, and optional AI chat.
View source on GitHubLive docs
Stack: Python, React, FastAPI, NSE, BSE, intraday, backtesting, technical analysis, AI agents, MIT
What it is
A full-stack market workspace for NSE and 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
- Intraday scanner — seven IST-aware modes: pre-market, pre-open, live scan, last hour, post-market review, single symbol, and weekend planning
- Dashboards — market overview, index and sector heatmaps, equity views, persisted backtest runs
- Strategy workbench — deterministic strategies across equities, indices, ETFs, and more
- AI-assisted chat — tool-first, source-aware, with guardrails against unsupported numeric claims
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.
At a glance
| Major surfaces | 4 |
|---|---|
| Intraday modes | 7 |
| Docs | Hosted live |
| Licence | MIT |
Not investment advice, and never a profit promise. It's a workspace for thinking about the market deterministically — the AI just helps you ask better questions of data you can already see.