# Kvants Whitepaper

## Kvants Whitepaper

- [Kvants Studio: The Quantitative AI Trading Agent Creator](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-quantitative-ai-trading-agent-creator.md): A self-directed software studio for building quantitative AI trading agents from transparent models, validation, education, and export tooling.
- [The Problem with News-Trading Agents](https://kvants.gitbook.io/kvants-whitepaper/the-problem-with-news-trading-agents.md): Why headline-reactive agents need quantitative models, sizing, validation, and risk structure before they can be trusted as software.
- [Platform at a Glance](https://kvants.gitbook.io/kvants-whitepaper/platform-at-a-glance.md): A map of the logged-in surfaces and how each one fits the self-directed workflow.
- [Home Command Center](https://kvants.gitbook.io/kvants-whitepaper/platform-at-a-glance/home-command-center.md): The logged-in hub for strategies, paper states, workspace context, and quick navigation.
- [Guided Onboarding](https://kvants.gitbook.io/kvants-whitepaper/platform-at-a-glance/guided-onboarding.md): A stepwise path from first idea to reviewed strategy, export, and paper workflow.
- [Kvants Studio: The Strategy Canvas](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas.md): The visual graph builder where quantitative logic becomes an editable strategy.
- [Node Graph Canvas](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas/node-graph-canvas.md): The strategy graph is the transparent model layer inside the agent.
- [Node Catalog](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas/node-catalog.md): The verified catalog of 71 node definitions across eight categories.
- [Templates](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas/templates.md): Sixteen starting points for common strategy shapes, ready to inspect and adapt.
- [AI Authoring](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas/ai-authoring.md): Plain language assistance for generating, modifying, and explaining strategy graphs.
- [Validation Gates](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas/validation-gates.md): The checks that keep validation visible before paper or export workflows.
- [Dual Backtest Engine](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas/dual-backtest-engine.md): NautilusTrader plus internal vector engines for validation and signal research.
- [Exports](https://kvants.gitbook.io/kvants-whitepaper/kvants-studio-the-strategy-canvas/exports.md): Portable Pine Script v6 and .kvants.json graph exports for user-controlled workflows.
- [The Signal Library](https://kvants.gitbook.io/kvants-whitepaper/the-signal-library.md): A catalog of quantitative signal families with evidence labels and research context.
- [Research-Backed Strategy Library](https://kvants.gitbook.io/kvants-whitepaper/the-signal-library/research-backed-strategy-library.md)
- [The Signals](https://kvants.gitbook.io/kvants-whitepaper/the-signal-library/the-signals.md): Twenty five pure signals grouped by trend, mean reversion, breakout, and volume families.
- [Using a Signal](https://kvants.gitbook.io/kvants-whitepaper/the-signal-library/using-a-signal.md): How a signal becomes one part of a broader agent with sizing, filters, and risk controls.
- [The AI Co-Pilot and the Strategy Brain](https://kvants.gitbook.io/kvants-whitepaper/the-ai-co-pilot-and-the-strategy-brain.md): AI assistance for authoring, explanation, diagnostics, and strategy memory.
- [Co-Pilot Chat](https://kvants.gitbook.io/kvants-whitepaper/the-ai-co-pilot-and-the-strategy-brain/co-pilot-chat.md): Generate, modify, and explain strategy logic from user instructions.
- [Strategy Brain](https://kvants.gitbook.io/kvants-whitepaper/the-ai-co-pilot-and-the-strategy-brain/strategy-brain.md): Identity, regimes, calibration, edge profile, failures, and memory panels for each strategy.
- [Red Team and Conscious Cycle](https://kvants.gitbook.io/kvants-whitepaper/the-ai-co-pilot-and-the-strategy-brain/red-team-and-conscious-cycle.md): Adversarial checks and deeper review loops for strategies that need more scrutiny.
- [Proof: Backtesting and Validation](https://kvants.gitbook.io/kvants-whitepaper/proof-backtesting-and-validation.md): Simulated testing, diagnostics, and stress workflows for strategy review.
- [Strategy Lab](https://kvants.gitbook.io/kvants-whitepaper/proof-backtesting-and-validation/strategy-lab.md): Run, compare, and revise strategies against historical data.
- [Performance Analytics](https://kvants.gitbook.io/kvants-whitepaper/proof-backtesting-and-validation/performance-analytics.md): Equity curves, metrics, and heatmap style diagnostics with explicit simulated labeling.
- [Stress and Walk Forward](https://kvants.gitbook.io/kvants-whitepaper/proof-backtesting-and-validation/stress-and-walk-forward.md): Out-of-sample folds, crisis windows, and cost realism for model resilience checks.
- [The AI Trading Terminal](https://kvants.gitbook.io/kvants-whitepaper/the-ai-trading-terminal.md): A research cockpit where charts, agent context, and paper controls sit together.
- [Terminal](https://kvants.gitbook.io/kvants-whitepaper/the-ai-trading-terminal/terminal.md): Chart-native agent research with paper and confirm-gated actions.
- [Terminal Pro](https://kvants.gitbook.io/kvants-whitepaper/the-ai-trading-terminal/terminal-pro.md): A denser strategy cockpit with paper deployment controls and telemetry panels.
- [The AI Multi-Strategy Builder](https://kvants.gitbook.io/kvants-whitepaper/the-ai-multi-strategy-builder.md): A software workflow for composing several quantitative agents into one research allocation.
- [Builder](https://kvants.gitbook.io/kvants-whitepaper/the-ai-multi-strategy-builder/builder.md): Combine strategy components and review allocation logic as software.
- [Wizard](https://kvants.gitbook.io/kvants-whitepaper/the-ai-multi-strategy-builder/wizard.md): A guided flow for choosing strategy families, constraints, and review steps.
- [User-Owned Execution](https://kvants.gitbook.io/kvants-whitepaper/user-owned-execution.md): The execution boundary: Kvants is research and export software; the user operates any live runner.
- [Paper Deployments](https://kvants.gitbook.io/kvants-whitepaper/user-owned-execution/paper-deployments.md): Hosted simulation and paper observation without Kvants placing real orders.
- [Export Package](https://kvants.gitbook.io/kvants-whitepaper/user-owned-execution/export-package.md): The target frozen package for moving strategy logic into user-owned infrastructure.
- [Self-Hosted Runner](https://kvants.gitbook.io/kvants-whitepaper/user-owned-execution/self-hosted-runner.md): A planned Railway template deployed and controlled by the user.
- [What Kvants Never Touches](https://kvants.gitbook.io/kvants-whitepaper/user-owned-execution/what-kvants-never-touches.md): Secrets, order history, logs, and runtime controls stay outside Kvants.
- [Stocks Research and Universes](https://kvants.gitbook.io/kvants-whitepaper/stocks-research-and-universes.md): Equities research surfaces, Alpaca-linked universe tools, and paper-oriented review.
- [Stocks Dashboard](https://kvants.gitbook.io/kvants-whitepaper/stocks-research-and-universes/stocks-dashboard.md): A dashboard for Alpaca-linked equities research context.
- [Stock Universes](https://kvants.gitbook.io/kvants-whitepaper/stocks-research-and-universes/stock-universes.md): Universe browsing and sync from the user's connected Alpaca context.
- [The MCP Agent Surface](https://kvants.gitbook.io/kvants-whitepaper/the-mcp-agent-surface.md): A programmable research surface for building, backtesting, inspecting, and analyzing strategies.
- [MCP Access](https://kvants.gitbook.io/kvants-whitepaper/the-mcp-agent-surface/mcp-access.md): Connections, scopes, and metered tool access for programmable research workflows.
- [Tools and Resources](https://kvants.gitbook.io/kvants-whitepaper/the-mcp-agent-surface/tools-and-resources.md): Sixty three tools, thirteen resources, and six prompts verified in the MCP server.
- [The Strategy Marketplace](https://kvants.gitbook.io/kvants-whitepaper/the-strategy-marketplace.md): A place to browse, publish, and copy strategy blueprints as generic research examples.
- [Discover](https://kvants.gitbook.io/kvants-whitepaper/the-strategy-marketplace/discover.md): Browse public strategy blueprints and copy graph logic for research.
- [Publish](https://kvants.gitbook.io/kvants-whitepaper/the-strategy-marketplace/publish.md): Submit a strategy listing with provenance and stale-result guards.
- [Leaderboard and Competition Arena](https://kvants.gitbook.io/kvants-whitepaper/leaderboard-and-competition-arena.md): Simulation-based rankings and contests for comparing strategy work.
- [Leaderboard](https://kvants.gitbook.io/kvants-whitepaper/leaderboard-and-competition-arena/leaderboard.md): Rankings based on persisted simulated metrics, with clear limits.
- [Competition Arena](https://kvants.gitbook.io/kvants-whitepaper/leaderboard-and-competition-arena/competition-arena.md): Contest workflows with anti-gaming rules and simulated scoring.
- [The Academy, Practice, and the Journal](https://kvants.gitbook.io/kvants-whitepaper/the-academy-practice-and-the-journal.md): Learning paths, planted-edge practice, and journal memory for strategy improvement.
- [Worlds and Lessons](https://kvants.gitbook.io/kvants-whitepaper/the-academy-practice-and-the-journal/worlds-and-lessons.md): Forty seven Playbook lessons, eighteen Coach lessons, and four built worlds plus a future primer.
- [Practice](https://kvants.gitbook.io/kvants-whitepaper/the-academy-practice-and-the-journal/practice.md): Deterministic planted-edge challenges and practice quests for model-building skill.
- [Journal](https://kvants.gitbook.io/kvants-whitepaper/the-academy-practice-and-the-journal/journal.md): A strategy memory system for hypotheses, gates, and lessons learned.
- [AI Credits: The KVAI Utility Token](https://kvants.gitbook.io/kvants-whitepaper/ai-credits-the-kvai-utility-token.md): KVAI is the consumption token for metered AI compute inside Kvants Studio.
- [Metered Engine](https://kvants.gitbook.io/kvants-whitepaper/ai-credits-the-kvai-utility-token/metered-engine.md): Hold, generate, settle actual metered cost, and burn what is consumed.
- [Effort Levels](https://kvants.gitbook.io/kvants-whitepaper/ai-credits-the-kvai-utility-token/effort-levels.md): LOW to MAX effort controls determine how much reasoning, output, tools, and refinement a request may use.
- [VIP Access Tiers](https://kvants.gitbook.io/kvants-whitepaper/ai-credits-the-kvai-utility-token/vip-access-tiers.md): Staked KVAI unlocks VIP tiers from VIP0 to VIP4 and platform feature access.
- [Getting KVAI](https://kvants.gitbook.io/kvants-whitepaper/ai-credits-the-kvai-utility-token/getting-kvai.md): Users acquire KVAI at their own initiative and deposit it for platform usage.
- [Ledger and Burns](https://kvants.gitbook.io/kvants-whitepaper/ai-credits-the-kvai-utility-token/ledger-and-burns.md): Off-chain ledger entries record deposits, holds, releases, and operational compute settlement.
- [Token Design](https://kvants.gitbook.io/kvants-whitepaper/ai-credits-the-kvai-utility-token/token-design.md): Utility-design facts: purchased for use, usable in the platform, and consumed by AI compute.
- [Affiliate and Growth](https://kvants.gitbook.io/kvants-whitepaper/affiliate-and-growth.md): Referral attribution and ecosystem growth loops for software adoption, without execution-linked compensation claims.
- [Architecture and Technology](https://kvants.gitbook.io/kvants-whitepaper/architecture-and-technology.md): The technical stack, product boundaries, and honest feature-status labels.
- [Roadmap](https://kvants.gitbook.io/kvants-whitepaper/roadmap.md): Forward-looking work items with status labels and no outcome promises.
- [Legal, Regulatory Positioning, and Risk](https://kvants.gitbook.io/kvants-whitepaper/legal-regulatory-positioning-and-risk.md): The required legal notice, user responsibility language, software boundary, KVAI utility framing, and risk statements.