AI Portfolio Analysis
Analyze investments, suggest any re-allocations to help in making better and smarter decisions.
Upload a CAS or CSV, or just describe your holdings. Portfolyze goes beyond basic tracking—it computes your real XIRR, suggests any re-allocations, and gives you AI-driven Hold, Sell, or Book Profits verdicts to help you make better and smarter investment decisions.
Sample analysis
Illustrative data
Total Value
₹18,42,600
Portfolio XIRR
16.8%
Benchmark (Nifty 50)
12.1%
Alpha
+4.7%
Beating the benchmark by 4.7 points this year
In short
Portfolyze is a web-based portfolio analysis tool for Indian investors. Upload a CAS PDF or CSV, or describe your holdings in plain English, and it computes your real XIRR, benchmarks it against the Nifty 50, suggests re-allocations and gives a Hold, Sell or Book Profits verdict on each holding. Manual tracking is free; the AI analysis is the Pro tier.
What can I upload to Portfolyze?
A Consolidated Account Statement (CAS) PDF or a CSV export from your broker or fund house. You can also describe your holdings in plain English.
Is Portfolyze free?
Manual holdings entry, basic value tracking and the Nifty 50 benchmark chart are free. CAS and CSV upload, AI parsing, XIRR and the verdicts are in Pro.
Is my portfolio data kept private?
Analysis is session-only by default, your data is never sold or shared with third parties, and market data comes only from public sources: NSE, BSE and AMFI.
Do I need to install an app?
No. Portfolyze runs in the browser, with no continuous syncing and no mobile app download.
- Web-based engine — no continuous syncing or mobile app downloads required. Session-only by default.
- Never sold or shared with third parties
- Public market data only — NSE, BSE, AMFI
How it works
From a raw account statement to a verdict on every holding — in under a minute.
-
Enter your holdingsUpload a CAS PDF, a CSV, or just type your portfolio in plain English.
-
Deep investment analysis & re-allocation suggestionsWe analyze your investments and calculate real money-weighted XIRR, benchmarked against the Nifty 50, while suggesting any re-allocations.
-
Get a verdict on every assetHold, Sell, or Book Profits — explained in plain language, not jargon.
-
See what you're leaving behindOpportunity-cost comparisons against sector-leading peers, plus a forecast.
What's inside
Everything you need to judge a portfolio the way an analyst would.
-
Plain-English entryDescribe your holdings in a sentence — no rigid templates, no manual data entry.
-
CAS & CSV importPull your holdings straight from your Consolidated Account Statement or a spreadsheet.
-
Deep Investment AnalysisReal money-weighted XIRR and historical trends, never approximated.
-
Smarter Investment DecisionsAI-driven Hold, Sell, or Book Profits verdicts based on mathematical data.
-
Asset Re-allocation SuggestionsIdentify sector or market-cap overexposures and get actionable re-allocation suggestions.
-
A downloadable PDF reportA complete, print-ready summary of your portfolio, verdicts, and forecasts.
See it before you use it
A sample analysis — illustrative data only.
Total Value
₹18,42,600
Portfolio XIRR
16.8%
Benchmark
12.1%
Alpha
+4.7%
HDFC Bank
Hold
Outperforming sector peers with strong fundamentals — continue holding.
+18.2%
The Portfolyze Forecasting Mechanics
Transparent, mathematical AI models powering your portfolio's future.
Core Forecasting Concepts
The ARIMA Model Explained
At the heart of our short-term predictions lies the ARIMA (AutoRegressive Integrated Moving Average) model, a cornerstone of econometric time-series forecasting. Financial markets are inherently non-stationary—they exhibit unpredictable trends and shifting volatility. ARIMA tackles this by breaking down the time series into three distinct mathematical parameters (p, d, q):
• AutoRegressive (p=1): Captures the relationship between the current asset return and its immediate past. By analyzing the lag, it detects residual momentum or immediate mean-reversion behavior in the market.
• Integrated (d=0): In traditional price forecasting, raw prices must be differenced to achieve stationarity. Because Portfolyze operates directly on logarithmic returns—which are mathematically stable over time—our model strictly uses an order of 0, effectively operating as an optimized ARMA engine.
• Moving Average (q=1): Accounts for the dependency between an observation and a residual error from a moving average model. It filters out random "white noise" and volatility shocks, isolating the true underlying signal.
By fitting this (1, 0, 1) model, Portfolyze maximizes prediction efficiency for standard financial forecasting, providing highly accurate momentum projections for the first 6 months. It is constrained only by our Dynamic Volatility Cap, capping extreme growth at 1.5 standard deviations above the market average.
• AutoRegressive (p=1): Captures the relationship between the current asset return and its immediate past. By analyzing the lag, it detects residual momentum or immediate mean-reversion behavior in the market.
• Integrated (d=0): In traditional price forecasting, raw prices must be differenced to achieve stationarity. Because Portfolyze operates directly on logarithmic returns—which are mathematically stable over time—our model strictly uses an order of 0, effectively operating as an optimized ARMA engine.
• Moving Average (q=1): Accounts for the dependency between an observation and a residual error from a moving average model. It filters out random "white noise" and volatility shocks, isolating the true underlying signal.
By fitting this (1, 0, 1) model, Portfolyze maximizes prediction efficiency for standard financial forecasting, providing highly accurate momentum projections for the first 6 months. It is constrained only by our Dynamic Volatility Cap, capping extreme growth at 1.5 standard deviations above the market average.
Gravity Blend and Linear Decay
In physics and spatial economics, the Gravity Model predicts interactions based on the "mass" of two bodies and the "distance" between them. Portfolyze applies a proprietary financial adaptation of this known as the Gravity Blend.
In financial markets, a fundamental law is "mean reversion"—an asset's tendency to eventually return to its historical equilibrium. The long-term macroeconomic average acts as a massive gravitational body (the "Mass"), represented in our engine by the 11% Capital Market Assumption (CMA).
Starting at Month 7 of our forecast, we introduce Linear Decay across the dimension of time (the "Distance"). As the forecast stretches further into the future, the gravitational pull of the market mean becomes mathematically dominant over short-term ARIMA momentum. This decay gently pulls the asset's predicted growth rate down from its extreme highs (or up from its lows). By Month 24, the influence of short-term volatility is fully nullified, and the prediction lands smoothly at the 11% long-term equilibrium.
In financial markets, a fundamental law is "mean reversion"—an asset's tendency to eventually return to its historical equilibrium. The long-term macroeconomic average acts as a massive gravitational body (the "Mass"), represented in our engine by the 11% Capital Market Assumption (CMA).
Starting at Month 7 of our forecast, we introduce Linear Decay across the dimension of time (the "Distance"). As the forecast stretches further into the future, the gravitational pull of the market mean becomes mathematically dominant over short-term ARIMA momentum. This decay gently pulls the asset's predicted growth rate down from its extreme highs (or up from its lows). By Month 24, the influence of short-term volatility is fully nullified, and the prediction lands smoothly at the 11% long-term equilibrium.
The IPO Fallback Protocol
Time-series AI models require statistically significant datasets to execute maximum likelihood estimations. Newly listed companies (IPOs) or freshly launched mutual funds are historically volatile and exhibit extreme standard deviations. If an asset has less than 48 months (4 years) of trading history, attempting to fit an ARIMA(1,0,1) model will result in high-variance statistical hallucinations—the sample size is simply too small to separate signal from noise.
To prevent model errors from distorting the reporting and driving misguided decisions, our IPO Fallback protocol acts as a quantitative safety net. If the 48-month data threshold isn't met, Portfolyze automatically bypasses the ARIMA engine entirely and assigns the asset a conservative, flat 11% growth curve (our Capital Market Assumption).
To prevent model errors from distorting the reporting and driving misguided decisions, our IPO Fallback protocol acts as a quantitative safety net. If the 48-month data threshold isn't met, Portfolyze automatically bypasses the ARIMA engine entirely and assigns the asset a conservative, flat 11% growth curve (our Capital Market Assumption).
Applying the Mechanics (Horizon: 24 Months | Data: 5 Years)
How We Forecast Stocks
Our stock forecasting engine looks at the daily closing prices of individual equities.
• Data Input: The last 60 months of market closing prices.
• Months 1-6: 100% ARIMA momentum, constrained by the Dynamic Volatility Cap.
• Months 7-24: Gravity Blend (Linear Decay) pulls the forecast toward the 11% average.
• Safety: IPO Fallback is triggered for stocks younger than 4 years.
• Data Input: The last 60 months of market closing prices.
• Months 1-6: 100% ARIMA momentum, constrained by the Dynamic Volatility Cap.
• Months 7-24: Gravity Blend (Linear Decay) pulls the forecast toward the 11% average.
• Safety: IPO Fallback is triggered for stocks younger than 4 years.
How We Forecast Mutual Funds
While the math remains identical, mutual funds are built on a different data foundation.
• Data Input: We pull the last 60 months of historical NAVs (Net Asset Values) rather than stock prices.
• Peer Comparison: Funds are compared against other funds in their exact category (e.g., Large Cap) ensuring comparisons are mathematically sound.
• Execution: The exact same 24-month horizon applies (ARIMA, Gravity Blend, IPO Fallback).
• Data Input: We pull the last 60 months of historical NAVs (Net Asset Values) rather than stock prices.
• Peer Comparison: Funds are compared against other funds in their exact category (e.g., Large Cap) ensuring comparisons are mathematically sound.
• Execution: The exact same 24-month horizon applies (ARIMA, Gravity Blend, IPO Fallback).
Opportunity-Cost & Peer Comparison Strategies
Comparing Against the Best (Top 1 vs. Top 3 Blend)
A forecast is only useful when compared against actionable alternatives. Portfolyze doesn't just predict how your current asset will perform—it compares it side-by-side with the sector's leading peers.
• Top Single Performer (Top 1): We project your initial capital into the #1 performing peer in the exact same category, letting you see the maximum absolute opportunity cost of your current holding.
• Top 3 Even Split (Blend): Because betting on a single winner carries concentration risk, our engine simulates a diversified alternative: an even 33% split across the Top 3 performers in the category. We calculate the individual ARIMA and Gravity Blend curves for all three peers and mathematically merge them into a single, risk-adjusted alternative trajectory.
• Top Single Performer (Top 1): We project your initial capital into the #1 performing peer in the exact same category, letting you see the maximum absolute opportunity cost of your current holding.
• Top 3 Even Split (Blend): Because betting on a single winner carries concentration risk, our engine simulates a diversified alternative: an even 33% split across the Top 3 performers in the category. We calculate the individual ARIMA and Gravity Blend curves for all three peers and mathematically merge them into a single, risk-adjusted alternative trajectory.
Interactive Mix Dropdown & Rebalancing
Within the Portfolyze dashboard, you have full control over how you visualize these alternatives via the Mix Dropdown.
You can instantly toggle between retaining your Current Mix, swapping out laggards for the Top 1 Peer, or reallocating into the safer Top 3 Blend. The AI dynamically recalculates your entire portfolio's future projected value and overall XIRR, allowing you to visually weigh the risk vs. reward of rebalancing before you make a single trade.
You can instantly toggle between retaining your Current Mix, swapping out laggards for the Top 1 Peer, or reallocating into the safer Top 3 Blend. The AI dynamically recalculates your entire portfolio's future projected value and overall XIRR, allowing you to visually weigh the risk vs. reward of rebalancing before you make a single trade.
The Intelligence Behind the AI Verdicts
Portfolyze evaluates both Historical 6-Month Momentum and Future ARIMA Projected Trajectory simultaneously. When opportunity costs exceed 20%, our AI runs a rigorous 4-quadrant friction matrix to deliver highly nuanced advice:
Sell (Switch) - Strong Switch
History & Future Align: Both recent 6-month performance and advanced future projections show the peer compounding at a significantly faster rate. Switch immediately to optimize trajectory.
Hold (Recent) / Switch (Projected) - Future Breakout
Future Breakout: Recent 6-month momentum has been identical, giving no immediate reason to panic. However, advanced ARIMA models detect long-term strength and project the peer to pull away significantly in the future. Switch if you are optimizing for a 24-month horizon.
Hold (Wait & Watch) - False Positive
False Positive: The peer had great momentum over the last 6 months, but our future projections indicate the growth curves are converging. Do not chase short-term momentum; wait to see if the asset recovers before incurring exit loads.
Hold - Permanent Hold
Flat Trajectory: While the peer generated massive returns years ago, both recent history and future projections show identical flat growth angles. Switching now incurs taxes with zero forward-looking advantage.
Sell (Preservation)
Toxic Asset: Triggered independently of peers when an asset is fundamentally broken or highly distressed (e.g., toxic penny stocks). Exit immediately to stop the bleeding.
Book Profits
Mean Reversion: Triggered when short-term gains heavily deviate from historical moving averages. Lock in gains before gravity and mean-reversion pull the price back down.
Built to keep your data yours
Your holdings are yours. Here's how Portfolyze treats them.
-
Session-only by defaultYour portfolio data lives in memory for your session. Nothing is stored unless you choose to export it.
-
No brokerage login requiredEnter your holdings directly. Portfolyze never asks for your broker or demat credentials.
-
Public market data onlyPricing and benchmarks are sourced from NSE, BSE, and AMFI — never sold or shared with third-party analytics.
See what your portfolio is really doing.
Two minutes, no signup required to see a sample.
Analyze My Portfolio →