of individual F&O traders finish in loss.
Chennai · built on Google Cloud, asia-south1
The market-intelligence terminal for Indian equities and derivatives.
Live market breadth, thematic money-flow, proprietary momentum signals, order-block structure, and an AI layer that narrates the session while you trade it.
128Features shipped
~300Symbols streaming live
~459Commits, 15 months
Oct 2026Paid launch target
Live on Google Cloud Run · trial cohort onboarded · pre-revenue today.
01Retail sees price. Institutions see everything else.
02
An entire market of
The problem
An entire market of
first-generation traders,
losing.
SEBI's own study of individual trading in equity F&O put numbers on what the industry preferred to leave vague.
aggregate retail net losses across FY22–FY24.
average loss per trader, per year.
Source: SEBI research on individual trading in equity F&O, September 2024. Figures last verified —.
03
Three failures, none of
Why it happens
Three failures, none of
them about effort.
Breadth turns before price
Market breadth and thematic money-flow turn before price does. Retail platforms show neither, so the regime change is invisible until the candle has already printed it.
The options math is hidden
Every mainstream Indian backtest models option strategies on the underlying. Theta and convexity are fiction, and the difference is discovered live, with real money.
No feedback loop
Traders never learn under which market conditions their own trades win. Without that, every loss is noise rather than information.
The individual sees candles. Institutions see everything else.
04
Three moats, each one
Defensibility
Three moats, each one
expensive to copy.
A competitor clones the interface in a month. The infrastructure underneath is the actual product.
Moat 01
Indicators
India-native proprietary indicators
Two branded numbers, both computed on full-universe live tick data, both unavailable anywhere else in the Indian retail stack.
- VARS — Volatility-Adjusted Relative Strength. Relative strength against NIFTY, normalised by ATR, so high-beta and low-beta names are directly comparable. Computed on five windows (5m / 15m / 25m / 60m / daily) at symbol, sub-theme and theme level.
- Anchored Market Breadth — breadth measured against rolling five-minute anchor baselines instead of the day's open, so an intraday regime turn is visible while day-open breadth still looks healthy. The family includes the Momentum Anchor Finder, a theme-count participation gauge and multi-baseline collapse detection used for exits.
Moat 02
Backtesting
Backtests that fill at real premiums
NIFTY weekly ITM option one-minute premiums are captured nightly, so a simulated option trade fills at the premium that actually existed — theta and convexity included.
- Walk-forward, no-lookahead computation contracts.
- Hash-versioned reproducible runs — strategy config hash, detection config hash, date range and trade count.
- A fill-audit tool that catches fantasy fills against each instrument's actual daily range.
- Whole-app replay of any historical trading day through the live terminal.
Moat 03
AI-native
Built for LLM clients, today
MCP turned "hand an LLM your broker keys" into a weekend project. Every one of those bots has a brain and hands, and is blind to everything in between that actually decides the trade.
- A remote MCP server exposing themes, scores, setups, positions and the backtest corpus to any LLM client — 18 tools, OAuth 2.1, billing caps.
- A five-minute AI market monitor and a voice narrator that speaks regime changes aloud.
- A one-prompt, client-grade weekly intelligence report.
- A good/bad rating loop on every AI alert, building a calibration corpus.
05
Proof of build
Not a deck. A codebase.
Five repositories over one shared BigQuery dataset — zenith-aerts.aerts, asia-south1. Fifteen months, roughly 459 commits across the three core repos.
| Repository | Role | Substance |
|---|---|---|
| aerts | Data foundation | Nightly ETL, indicator computation, scoring and options capture on a GCE VM driven by cron. First commit 19 May 2025 · 217 commits. |
| aertsv2 | Production terminal & backtesting IP | 187 routes · 5 simulators · 24 YAML strategy configs · 8 backtest tables. |
| aertsv3 | Clean-architecture rewrite | Flask + Socket.IO · 115 routes · 42k Python LOC · 1,338 test functions. Carved out of v2; first commit 21 Jul 2026. |
| aerts-mcp-server | Remote MCP layer | 18 tools · 76 tests · OAuth 2.1 · billing caps. Read-only analytics surface. |
| aerts-studio | Content pipeline | Turns the weekly intelligence report into narrated video. Joined to the platform only through a BigQuery table. |
Commit counts and route counts read directly from the repositories. Last verified —.
06
The product
128 features, in ten areas.
Eighteen modules in the live terminal
01 Live Theme Board
02 Theme Deep-Dive
03 Symbol Workstation
04 One-Click Futures
05 Five-Tier Breadth
06 SMC Structure
07 Live Alerts Feed
08 OBOEE
09 PVOEE
10 AI Market Monitor
11 Automated Exit Engine
12 Live Positions & P&L
13 Configurable Gating
14 Signal Board
15 Theme Overview
16 Theme Universe Builder
17 Backtest Analytics
18 Trade Replay
Feature split by area
| Area | # |
|---|---|
| Backtesting & simulation | 21 |
| Data, alerts, journal, platform | 17 |
| AI & research layer | 14 |
| Thematic engine | 13 |
| Scoring & indicators | 12 |
| Execution bots (signal engines) | 12 |
| Live terminal | 11 |
| SMC / order blocks | 11 |
| Market breadth suite | 9 |
| Options tooling | 8 |
| Total | 128 |
Screenshots are from the live AERTS terminal. Figures shown are illustrative of the interface, not a record of returns.
07
Compliant by architecture,
Regulatory posture
Compliant by architecture,
not by promise.
SEBI's February 2025 algo-trading framework requires broker empanelment and human-in-the-loop for third-party execution. The system was built that way from the start.
- The MCP server's analytics surface is read-only, with zero access to order execution — no broker API, no execution credentials, no connection to the live trading systems.
- The execution rail lives separately, in aertsv3, behind one-click manual confirm.
- Order credentials are the acting user's own, with no fallback — a failed credential lookup fails the order rather than quietly placing it under a box identity.
- The signal engines — OBOEE, MOEE, NRF, PVOEE, DEQ and the automated exit engine — surface setups for manual confirmation. They are not auto-execution for third parties.
Supporting precedent: Zerodha's own Kite MCP ships read-only — LLM order placement is deliberately excluded. India's largest broker will not let a language model place orders, which is precisely the argument for a compliant middle layer.
08
Everyone sells tools.
Market & whitespace
Everyone sells tools.
Nobody sells intelligence.
India leads the world in derivatives by contracts traded. Charting is commoditised. Execution is commoditised. Market intelligence is not served at all.
| Segment | Players | The gap |
|---|---|---|
| Options analytics | Sensibull, Quantsapp, Opstra | Traders demonstrably pay for this. Zero market context: no breadth, no rotation, no regime. |
| Algo execution | Streak, AlgoTest, Tradetron | Fills modelled on the underlying, with AlgoTest excepted. Intervention is confirm-or-kill on a black box. |
| EOD analytics | Trendlyne, StockEdge | Scores and screeners on yesterday's close. The regime turn stays invisible until tomorrow. |
| Charting | TradingView | The screen-time incumbent. AERTS is the intelligence on the second screen. |
| AERTS | — | Live breadth · thematic rotation · honest options backtests · AI-native. |
Market structure and competitor positioning as at August 2026. Every competitor in this table is broker-attached or broker-funded; none ships live market intelligence. Last verified —.
09
Business model
Four lines, one platform.
| Line | Price | What it is |
|---|---|---|
| Pro Terminal | ₹2,499/mo | The full workstation — breadth suite, VARS, themes, signals, journal. |
| Replay Academy | ₹999/mo | Learn on replayed market days with paper orders. Education revenue with zero additional build. |
| RIA White-Label | ₹25,000/mo | Per firm. Advisories get the weekly AI intelligence report and dashboards under their own brand. |
| Data & API | Custom | VARS, anchored breadth and the MCP server licensed to brokers, platforms and AI copilots. |
Pricing is benchmarked against comparable Indian tools. Validation against the current trial cohort is underway — it is not yet proven. AERTS is pre-revenue today.
10
Shipped, in progress,
Status
Shipped, in progress,
and honestly labelled.
Shipped
- Live terminal on Cloud Run, ~300 symbols streaming
- 128 features across ten areas
- VARS and anchored breadth in production
- Premium-accurate options backtesting
- MCP server, 18 tools
- Voice narrator and AI market monitor
- Trial cohort onboarded
In progress
- Trial → paid conversion funnel
- Replay Academy launch
- Multi-broker integrations
- Pricing validation with the trial cohort
Roadmap
- Bot Studio — bring-your-own-strategy on AERTS data and rails
- Mobile companion
- Regional-language AI narration
- VARS and breadth licensing to brokers and platforms
- Partner-facing MCP and API
11
Zenith Tech Solutions,
The company
Zenith Tech Solutions,
Chennai.
A private limited company incorporated in 2022. AERTS is built and operated by its founder, Raaj Vinay — a single-founder team today, which is stated here rather than left for diligence to find.
DPIIT · Startup India recognised
StartupTN registered
Google Cloud · asia-south1
Incorporated 2022 · Chennai, TN
Start a conversation
Investor deck and
technical annexure,
on request.
The detailed materials — financial model, architecture annexure, backtest methodology and compliance memo — are shared directly rather than published. Tell us who you are and what you need to see.
For investors, grant and incubation programmes, brokers and advisory firms.
enable JavaScript to reveal Replies come from the founder directly. No mailing list, no drip sequence.