QuantOI
Systematic options-trading intelligence for Indian index derivatives
- Type
- In-house R&D Product
- Markets
- NSE / BSE index derivatives
- Status
- Paper-trading validation
- Interface
- Telegram bot
Our own R&D product: a system that watches live Nifty and Sensex options chains through the full trading day, separates genuine institutional conviction from noisy open-interest spikes, and turns confirmed signals into defined-risk spread trades — every one simulated through a realistic paper broker with real exchange cost modeling. It runs unattended and reports to a Telegram bot. Currently in disciplined paper-trading validation before any capital goes live.
The Problem
Raw open-interest and volume spikes — the signals most retail tools alert on — are frequently false moves ("stop-hunts"). Acting on them directly is how retail traders bleed money.
Naked option buying carries undefined risk, and most retail tooling looks at the options chain in isolation: no institutional positioning, no macro context, no cross-session persistence.
Testing a strategy honestly is itself an engineering problem: naive simulators ignore fill quality, order-size limits, and transaction costs — and quietly overstate results.
What We Built
We built the whole pipeline: real-time market data in, scored and filtered signals, risk-defined trade construction, realistic simulation, and a Telegram bot as the operations console — with a hard safety flag ensuring no path to live orders during validation.
Observation-window signal scoring
A candidate signal isn't trusted on arrival — it's scored over an observation window against OI persistence, volume confirmation, VWAP alignment, IV percentile, futures positioning, and higher-timeframe trend.
Defined-risk spread systems
Two strategies, both risk-capped by construction: directional debit spreads on confirmed breakouts, and credit spreads defending major open-interest walls. No naked exposure.
Realistic paper broker
Simulated fills respect order-book depth, exchange freeze-quantity order splitting, and the real F&O transaction cost structure — so results aren't quietly flattered.
Institutional context engine
SEBI participant-flow data, max pain, gamma walls, and cross-session OI persistence give every signal the context most retail tools ignore.
LLM-generated macro briefing
A daily pre-market briefing synthesized by an LLM from global markets and news, feeding sideways-day detection that suppresses trade suggestions in choppy conditions.
Telegram bot operations
Trade alerts, active position tracking, portfolio Greeks, institutional summaries, and system health — pushed automatically or on demand, from anywhere.
Self-learning calibration
Every signal outcome is logged into a monthly calibration pipeline that proposes — never silently applies — threshold adjustments as evidence accumulates.
Stock swing-trade screener
An independent daily module screens ~210 F&O stocks through Minervini's trend template, volatility-contraction detection, and a strict reward-risk gate — with near-misses shadow-tracked to keep thresholds honest.
The Outcome
- Runs unattended through every market session, logging roughly a thousand market observations a day into structured datasets.
- Every trade idea is costed and simulated realistically — order-book-aware fills and real exchange charges, not optimistic assumptions.
- Every outcome feeds the calibration loop, so the system's thresholds improve from evidence rather than intuition.
- A hard safety flag guarantees no execution path to real money until validation is complete — discipline first, capital later.
Interface visuals are illustrative mockups — client data is not shown.
What We Did
Technology
What's Next
- Cloud deployment so the system runs on managed infrastructure instead of a local machine.
- A web dashboard UI over the Telegram bot — positions, equity curve, signal history, and calibration reports in the browser.
- Live broker integration behind strict risk limits, once the paper track record has proven itself over a meaningful period.
- Productizing the alert engine so other systematic traders can subscribe to the intelligence layer.
Self-initiated R&D product, currently in paper-trading validation. No live-capital performance is claimed, and nothing here is investment advice.
Want something like this for your business?
Catalogs, lead capture, automation, AI tools — we design and build systems that do real work. Tell us about your problem and we'll propose a solution in plain language.
Start a Conversation