Build a modern, slick, dark-themed React single-page application for a product called Marketforge.ai.

The Product: An AI-powered market-making agent that provides automated, 24/7 liquidity across the long tail of prediction markets (like Kalshi). It monitors information sources, synthesizes signals, posts competitive bid/ask quotes, dynamically adjusts spreads based on information velocity, and pulls quotes to prevent adverse selection.

Tech Stack: React, Tailwind CSS, Framer Motion (motion/react), Recharts, and Lucide React icons.

Design System & Vibe:

• Aesthetic: Dark mode, professional, high-frequency trading terminal meets modern AI SaaS. Use deep dark backgrounds (e.g., #0a0a0a and #141414) with subtle white/indigo borders (border-white/10).
• Typography: Use Inter for standard UI text and JetBrains Mono for all numbers, financial data, and logs.
• Colors: Indigo for primary branding, Emerald for positive/active states, Rose for negative/pulled states, and Amber for warnings.

Application Structure:

The app should have a state-based router that switches between a Landing Page and a Dashboard.

1. Landing Page:

• Background: Dark with a subtle radial gradient and a faint CSS grid pattern.
• Hero Section: Animated using Framer Motion. Include a badge saying "Kalshi API Integration Live", a massive headline ("Provide Liquidity to the Prediction Economy"), a subheadline explaining the product, and a "Enter Dashboard" CTA button.
• Features Section: A 3-column grid highlighting: Continuous Synthesis, Dynamic Spreads, and Adverse Selection Protection.

2. Dashboard Layout:

• Sidebar: Navigation menu with tabs for "Active Markets", "Backtesting", and "Agent Settings", plus an "Exit Demo" button at the bottom.
• Topbar: Display a pulsing "AGENT: ACTIVE" status indicator, a "KALSHI API: CONNECTED" badge, a live "SESSION P&L" (e.g., +$1,240.50), and a master toggle button to "Halt Agent" / "Start Agent".

3. Dashboard Views (with realistic dummy data):

• Active Markets (Default): A split layout.
  • Left side: A list of market cards. Each card should show the Market ID, Category, Name, AI Confidence %, Spread, Bid, and Ask. Create different states for these cards: normal (green/red bid/ask), widened spread (yellow warning), and pulled quotes (dimmed out with a red warning about adverse selection).
  • Right side: A terminal-style "Agent Reasoning Log" that displays a scrolling feed of the AI's real-time actions (e.g., "Synthesized 42 new articles", "Spread tightened to 3¢", "PULLED QUOTES: Detected aggressive informed flow").
• Backtesting:
  • Top row: 4 stat cards showing Simulated P&L, Sharpe Ratio, Max Drawdown, and Win Rate.
  • Main area: A large, responsive Area Chart using recharts showing a simulated P&L curve over time with a nice indigo gradient fill.
• Agent Settings:
  • A section with checkboxes to toggle Market Categories (Politics, Macroeconomics, Entertainment, Sports).
  • A "Risk Parameters" section with HTML range sliders for: Max Position Size per Market, Base Spread Target, and Adverse Selection Sensitivity.