Fintola 2.0 – Live Stock & Strategy Dashboard

Overview

Fintola 2.0 is a real-time trading intelligence dashboard that helps active traders monitor live market movements, evaluate intraday and delivery (positional) strategies, and generate AI-assisted trade insights. It merges streaming price data, technical indicator computation, and Gemini-powered narrative analysis to support more informed decision-making. (Not financial advice.)

Core Capabilities

📈 Live Market Data

• Real-time price streaming for selected equities
• Bid/ask snapshots and last traded price movements
• High/low, % change, volume, VWAP (where available)
• Auto-refresh with throttled WebSocket / SSE handling

🧠 AI Insights (Gemini)

• Natural‑language summaries of market context
• Buy / sell inclination with reasoning (risk-weighted)
• Strategy alignment (intraday vs delivery)
• Sentiment & momentum interpretation
• Risk warnings and confidence scoring

🧪 Strategy Engine

• Intraday momentum & breakout model
• Delivery (swing / positional) trend confirmation
• Customizable indicator thresholds (e.g. RSI, EMA crossovers)
• Multi-timeframe confluence checks
• Risk/reward and stop-loss suggestion logic

📊 Technical Analysis

• Indicator overlays: EMA / SMA stacks, RSI, MACD, VWAP*
• Support / resistance zone detection*
• Candlestick pattern tagging*
• Volume profile & relative volume flags*
• Backtest skeleton for extending strategy validation (* where implemented or in progress)

🔔 Alerts & Signals

• Threshold & crossover triggers
• Strategy alignment signals (enter / watch / exit)
• Risk management prompts (trail / tighten / reduce)
• Debounced notification pipeline to avoid noise

Architecture

• Frontend: Next.js (App Router) + React server & client components
• Data Layer: WebSocket / SSE streaming + REST fallbacks
• Backend Services: API routes for symbol lookup, strategy evaluation, AI prompt orchestration
• Database: PostgreSQL via Prisma (watchlists, user preferences, cached indicator snapshots)
• AI Integration: Gemini API (structured prompts with system + context + constraints)
• Styling: Tailwind CSS + utility-driven responsive layout
• State Management: Lightweight client state + server components + SWR/React Query (if used)
• Caching: In-memory + incremental static regeneration (for non-live pages)

Data Flow

1. User selects / watches symbols
2. Stream handler ingests live ticks (normalized)
3. Indicator & strategy engine computes derived metrics
4. Snapshot persisted (batched) for historical context
5. AI prompt builder composes summary (symbol + indicators + strategy status)
6. Gemini returns structured JSON / text → rendered into UI panels
7. Alerts emitted when rules fire

Performance & Reliability

• Connection health monitoring & auto-reconnect
• Rate limiting + debounce for AI prompt invocations
• Batched DB writes to reduce churn
• Lazy loading heavy analysis panels
• Memoized indicator computations per symbol frame

Security & Compliance

• Environment-isolated API keys
• Input sanitization for AI prompt context
• Least-privilege DB access via Prisma schema scoping
• No PII storage (only user config + watchlists)

Challenges & Solutions

Real-time Stability

Managed reconnect storms with exponential backoff and sequence gap reconciliation.

Indicator Computation Load

Introduced caching layers and incremental recalculation (only on new ticks).

AI Hallucination Mitigation

Used structured prompts + explicit constraints + confidence extraction parsing.

Strategy Interpretability

Displayed raw indicator values alongside AI narrative for user trust.

Future Enhancements

• [ ] Full backtesting module with equity curve
• [ ] Multi-broker order placement integration
• [ ] Options chain analytics
• [ ] Portfolio risk aggregation
• [ ] Mobile-optimized condensed layout
• [ ] More advanced ML-based anomaly detection

Lessons Learned

Building Fintola 2.0 reinforced the importance of:

• Clear separation between raw market data and derived analytics
• Guardrails for AI output (never execute trades automatically)
• Performance budgeting for streaming UIs
• Transparent strategy logic to build user confidence

Not financial advice. Users should verify signals independently.