Press 1-4 to switch tabs · ⌘K to search · L learn mode · D dark mode
Market overview
Live index performance and ranked opportunity scan.
Connect your backend
Configure your backend URL in Settings to start exploring market data, forecasts, and portfolio simulations.
Learn mode
Reading the market overview
This page teaches how a model turns market data into scenarios. The ranking is not a buy list; it is a way to compare upside, downside, and confidence side by side.
Macro snapshot
Opportunity ranking
Base is the expected return under normal conditions. Bear and bull are downside
and upside scenario projections. Confidence reflects model certainty.
Market sentiment
--
/100
Waiting for market data
Key insights
Top opportunities (by base return)
Diagnostics and data health
View raw diagnostics
Health
Models
Refresh status
Ticker forecast
Historical price with bear, base, and bull scenario paths projected forward.
Understanding scenario paths
These are model-generated probability-weighted projections, not guarantees.
The base path reflects the central estimate while bear and bull paths show
downside and upside ranges.
Scenario path summary
Key insights
Portfolio simulator
Forecast-ranked asset allocation with bear, base, and bull portfolio values.
Allocation settings
Portfolio inputs
Set the amount, risk appetite, and horizon used by the portfolio classroom.
Portfolio classroom
How risk settings change the simulation
The simulator shows how asset mix, cash buffer, downside value, and confidence move when you change risk appetite or horizon. It is an educational sandbox, not a trade instruction.
Class allocation
Why hold cash?
The model reserves a cash buffer based on risk appetite and scenario balance.
Asset allocation (base scenario)
Simulation example trades (base scenario)
Key insights
Advanced diagnostics and backtest metrics
RL allocation summary
Backtest metrics
About
App context, system numbers, and builder profile.
Foresight is a full-stack market intelligence and financial literacy platform.
The app combines live market index history, market data refresh jobs, FastAPI inference routes, ML/RL allocation artifacts, scenario forecasting, backtests, explainability, and a polished frontend into one learning-oriented product.
Sudip Shrestha
Data Analyst & AI Engineer
I am a data enthusiast with a background in software engineering and a passion for transforming complex datasets into compelling visual stories. I am learning and building ML models.
System architecture
Data ingestion
- yfinance OHLCV for equities, ETFs, crypto, and market indices
- Configured universes from
config/asset_universe.v1.jsonandconfig/market_indices.v1.json - Macro CSV inputs for model context
Refresh and storage
- Render cron runs
scripts/refresh_supabase_daily.shfor daily Supabase updates - Supabase stores market history, profiles, macro context, index snapshots, and forecasts
- Market index snapshots can refresh on demand when the first cards are missing
FastAPI model layer
- Supabase forecast engine serves market overview, ticker profiles, forecasts, and simulations
- Local artifact engine remains available for optional RL diagnostics and backtests
- Index endpoints serve latest cards and historical chart series
Learning frontend
- Static HTML/CSS/JS app calls the API directly
- Market overview combines index history, scenario rankings, and sentiment
- Learn Mode turns outputs into plain-language finance lessons
Runtime contract
Render runs with Supabase required through SUPABASE_URL and
SUPABASE_SERVICE_ROLE_KEY. Local development can still enable artifact
fallback for offline diagnostics.
User-facing APIs
/api/market/indices, /api/market/indices/{symbol}/history,
/api/forecasts/ticker, /api/forecasts/market, and portfolio simulation
routes power the visible workflows.