Skip to main content
Talk to QuantLab the way you would talk to a research colleague. Below are ten example messages you can paste into chat. Adapt tickers, dates, and wording — you do not need to name tools or paste JSON configs.

1. What can you do?

Hi! What can you do with the QuantLab MCP https://quantx-api.limex.pro/mcps/quantlab ?

2. Multi-model stress on my holdings

That’s my current portfolio (ticker | amount in shares): AAPL | 42 AXP | 37 BA | 29 CAT | 8 GS | 7 HAL | 307 NVDA | 43 XLRE | 484 Tell me, how much will I lose tomorrow if Nvidia drops by 17%? Do not rely on a single stress-testing model. Apply multiple approaches: incorporate macroeconomic factors or model the portfolio with Nvidia as the core holding in vine copula. Here use TLT, SIZE and GLD as the factors. Use daily returns from 2024-07-09 to 2026-07-09 for all stress models — start and end only, no longer or shorter lookback. 1 day rebalance frequency. Try student t, vine copula and gaussian approach.

3. Sweep every optimizer, then advise

Evaluate my portfolio using all 73 portfolio optimization methods available in QuantLab: each method with each covariance estimation type. All portfolio optimization jobs at one time, no batching is needed. Based on each method, do I need to rebalance? Are any of my current positions significantly overweight? Additionally, I am hesitant to sell XLRE to increase my NVDA exposure given the current economic climate. Provide your recommendations.

4. Equity curve after a simple backtest

Run an equal-weight backtest on AAPL, MSFT, and GOOGL from 2023-01-01 through 2025-01-01, no machine learning. When it finishes, show me the portfolio equity curve and how each name’s position evolved over the evaluation window.

5. Price and feature charts

Pull Massive daily data for NVDA and AMD from 2024-01-01 to 2026-01-01. Plot NVDA’s close, then compute RSI and MACD and plot those features for NVDA on separate charts so I can compare the timing of signals.

6. Weight dynamics and a tearsheet

Take my last optimized portfolio and show how the weights moved over time. Also give me a full performance / risk tearsheet against SPY so I can see drawdowns and ratios in one place.

7. Pairwise dependence and clusters

On the same portfolio, plot the Gaussian copula density between NVDA and XLRE, and build a cluster dendrogram of all holdings — I want to see whether the REIT sleeve is really diversifying the tech names.

8. Factor impact story

Run a factor-impact analysis on my current weights using MTUM, VLUE, QUAL, and SIZE. Tell me in plain language which factors are driving the book and whether NVDA’s bet is mostly momentum or something else.

9. Classical ML forecasts into max Sharpe

For AAPL, MSFT, and GOOGL, use Massive data from 2022-01-01 to 2026-01-01. Engineer RSI and moving averages, train an XGBoost model with a moderate depth and a few hundred trees to forecast next-day returns, then optimize with max Sharpe using those predictions and backtest. Summarize whether the ML path beat equal weight over the held-out period.

10. LSTM path and a prediction chart

Same universe and dates as above, but use an LSTM instead of XGBoost. After training, show predicted versus realized test returns for AAPL, then run a min-vol portfolio on those forecasts and tell me if the equity curve looks smoother than the XGBoost max-Sharpe run.

Mix and match: paste a portfolio table, ask for charts, then follow up with stress or ML without repeating the holdings. QuantLab keeps the chain of results behind the scenes so you can stay in the conversation.