01
Market-aware research
Ask about NSE and BSE securities or switch to US research. Compute 1.5 keeps the active market in context, resolves tickers correctly, and uses the right currency and universe.
iStocks release / August 2026
Compute 1.5 is now the iStocks research engine.
This release expands the research surface from India to the US, makes price-target odds more explicit about their evidence, and connects discovery, charts, Sand, and strategy workflows in one system.
What changed
More of the stack is connected.
Compute 1.5 is a product release, not a promise of guaranteed returns. It gives the research agent more precise market context and more useful tools while keeping trading decisions transparent and reviewable.
02
Research skills that return data
The Trading Agent can find a stock, analyze it, screen a market universe, list movers, and query the market database instead of relying on a generic chat response.
03
US stock discovery
Screen and rank the S&P 500, inspect market movers, and research named US companies alongside the existing India-focused workflows.
04
Evidence-based target odds
For price-target questions, the analysis measures the required move and compares it with similar historical moves within the requested trading-session horizon.
05
Sand probability workspace
Krishna, Chanakya, and Aryabhata research historical evidence, current news, and numerical data independently before their findings are synthesized.
06
Chart control from a prompt
Request candlestick views, intraday or daily intervals, indicator overlays, and rule highlights directly from the Trading Agent conversation.
07
Strategy workflow depth
Move from an idea to a saved Sangraha strategy, backtest it on historical data, and use paper-trading controls before any broker execution. News-triggered rules stay controlled by their deployment settings.
Probability, made clearer
A probability is now tied to a method, not just a confident sentence.
Price-target analysis uses completed daily candles. It calculates the percentage move required, checks whether comparable upside or downside moves completed within the requested number of trading sessions, and reports the historical hit rate, sample size, recent average movement, and ATR context. Sand adds a separate multi-agent research verdict for broader market questions.
Input
Price, target, and trading-session horizon
Evidence
Historical comparable moves and volatility context
Output
Hit rate, odds band, samples, and caveats
Transition
Nothing to migrate manually.
Compute 1.5 is selected for new Trading Agent conversations. Requests created with the Compute 1.0 key resolve to Compute 1.5, so your existing sessions, quotas, and saved work continue without interruption.