Web interface

Run QuantAgent’s browser-based interface to select assets, configure timeframes, and read multi-agent trading analysis results.

Start the Flask development server from the project root:

python web_interface.py

The interface is available at http://127.0.0.1:5000 once the server is ready.

The web interface fetches all market data live from Yahoo Finance via yfinance. There is no support for uploading local CSV files through the UI. Ensure you have an active internet connection before running analysis.

Walkthrough

1. Select an asset

   Open the dropdown labelled Asset and choose one of the twelve built-in instruments or any custom asset you have added previously. The built-in options are:

   Code	Display name
   SPX	S&P 500
   BTC	Bitcoin
   GC	Gold Futures
   NQ	Nasdaq Futures
   CL	Crude Oil
   ES	E-mini S&P 500
   DJI	Dow Jones
   QQQ	Invesco QQQ Trust
   VIX	Volatility Index
   DXY	US Dollar Index
   AAPL	Apple Inc.
   TSLA	Tesla Inc.

2. Choose a timeframe

   Select the candle interval from the Timeframe dropdown. Each interval has a maximum lookback enforced by Yahoo Finance:

   Timeframe	Max days
   1m	7
   5m	60
   15m	60
   30m	60
   1h	730
   4h	730
   1d	730
   1w	730
   1mo	730

   The UI validates the date range against these limits and shows an error if you exceed them.

3. Set the date range

   Enter a Start date and End date using the date pickers. You can optionally pin the end time to the current moment by enabling Use current time, which is useful for live intraday analysis.

   Dates in the future are rejected automatically.

4. Run the analysis

   Click Analyze. QuantAgent fetches the OHLCV data from Yahoo Finance, slices the most recent 45 candles, generates candlestick and trend charts, and then passes everything through the four-agent pipeline (Indicator → Pattern → Trend → Decision).

   A loading indicator is shown while the agents are working. Analysis typically takes 15–60 seconds depending on your LLM provider’s latency.

5. Read the results

   Results are displayed on the output page with four sections:

   • Indicator report — RSI, MACD, Stochastic Oscillator, and other computed metrics with plain-language interpretation.
   • Pattern analysis — chart formation detected from the candlestick image (e.g. bull flag, golden cross).
   • Trend analysis — trend channel direction, slope, and consolidation zones derived from the annotated trend chart.
   • Final decision — a structured LONG or SHORT directive including the risk-reward ratio, forecast horizon, and concise justification.

   Both the candlestick chart and trend chart are rendered inline alongside their respective agent reports.

API key management

Your LLM provider API key can be set or changed at any time without restarting the server. Open the Settings panel in the UI and:

1. Select your active provider (OpenAI, Anthropic, or Qwen).
2. Paste your API key into the input field.
3. Click Save. The key is applied immediately to the running server via POST /api/update-api-key and validated with a lightweight test call.

The UI masks the stored key (showing only the first three and last three characters) and displays an inline error if the key is invalid, rate-limited, or associated with a billing issue.

You can also set keys before starting the server using environment variables — OPENAI_API_KEY, ANTHROPIC_API_KEY, or DASHSCOPE_API_KEY — so the Settings panel is pre-populated on first load.