Spawn AI Agents that
research alpha
vs NIFTY 50
+8.4%Sharpe
Win Rate
Max DD
Vol
Monthly P&L
Coverage
1,800+
Instruments 10 year daily data
Strategy Performance
Strategy Return
+15.3%
Benchmark
+6.3%
Sharpe Ratio
1.96
Max Drawdown
-4.1%
Smart Growth
12,847+
Strategies backtested on the platform, building research depth and reaching performance benchmarks.
1.2B
Data points analyzed across equities, derivatives, and indices for comprehensive research coverage.
+23.5%
Average annual return improvement across diversified strategies, balancing risk while steadily increasing growth.
1,800+
Stocks available to backtest across NSE and BSE, from leading indices to mid & small cap universe.
Smarter Research With AI Forecasts, Institutional Analytics, and Real-Time Data
Describe your strategy in plain English and get backtests within seconds, with zero unnecessary complexity.
RaptorBT is open-source, battle-tested, and blazing fast. Institutional-grade calculations with full transparency.
Get AI-powered regime detection, parameter sensitivity analysis, and Monte Carlo simulations to maximize long-term edge.
Sample Strategies
| Name | Return | Sharpe |
|---|---|---|
| 15min ORB | +18.7% | 2.24 |
| VWAP Scalp 5m | +12.3% | 1.89 |
| CPR Bounce | +16.1% | 2.05 |
| Momentum 5m | +14.8% | 1.73 |
| Supertrend 5m | +21.2% | 1.91 |
Need Help
Common questions about the platform, data coverage, and getting started with AI-powered quant research.
An AI agent is a specialized LLM with a focused tool surface. alphabench spawns four core agents — Planner (orchestration), Quant (backtest execution), Researcher (data discovery and fundamental screening), and Diagnostician (trade forensics) — that collaborate to handle your request. They share message history but each agent only sees the tools it owns, which keeps reasoning sharp and answers truthful.
The Planner agent reads your message and routes work. A backtest request goes to the Quant agent. A "find me stocks with ROE > 18" request goes to the Researcher. A "why did this strategy lose money" request goes to the Diagnostician. Most prompts use two or three agents in sequence — you see one coherent answer.
Describe your strategy in plain English — the Planner agent interprets it, the Quant agent translates it into a strategy DSL and calls RaptorBT (our open-source Rust engine), which runs the backtest deterministically and returns equity curves, trade logs, and metrics. No LLM is in the execution path, so backtests are reproducible.
The Quant agent handles any intraday strategy: Opening Range Breakouts (ORB), CPR levels, VWAP scalping, momentum breakouts, options straddles, futures trending. Use 1-minute to 15-minute timeframes with time-of-day filters, session-based exits, and real slippage modeling.
1,800+ NSE & BSE equities with 10+ years of daily data, active F&O instruments, weekly and monthly options chains, and 30+ fundamental fields per equity (ROE, ROCE, P/E, P/B, revenue growth, margins). Intraday data goes down to 1-minute candles. The Researcher agent has read-only access; the Quant agent uses the same data to backtest.
Yes. The Quant agent supports straddles, strangles, spreads, iron condors, calendar spreads, diagonals, and custom multi-leg strategies on NIFTY and BANKNIFTY with real historical options chain data and per-leg expiry control.
Yes. Paper trade with tick-level live simulation first, then deploy to live trading with Zerodha integration. Monitor P&L, positions, and signals in real-time via WebSocket updates. Strategy health monitoring runs as a separate agent loop after deployment.
Yes — you can start for free with no credit card required. The free tier includes generous daily backtest limits and full access to equity data.
Ready to Find Your Edge?
We're scaling our services. Request access and we'll notify you as soon as possible.
alphabench