diff --git a/BUILDING.md b/BUILDING.md new file mode 100644 index 00000000..c02d1e99 --- /dev/null +++ b/BUILDING.md @@ -0,0 +1,91 @@ +--- +last-updated: 2026-04-02 +--- + +# BUILDING + +## Timeline + +### Phase 1: MVP (Dec 2025) +- AI Chat interface with Claude (Haiku/Sonnet/Opus) +- Real-time market data via Alpaca (quotes, charts, WebSocket) +- Portfolio tracking with paper trading +- Options screener and strategy builder +- Emotional Firewall v1 (basic discipline prompts) +- Supabase auth + multi-tenant data +- Deployed to Vercel (deepstack.trade) + +### Phase 2: Research Platform (Dec 2025) +- Professional charts (Lightweight Charts, 30+ indicators) +- Thesis Engine (hypothesis tracking with live validation) +- Trade Journal (rich text, emotion tracking, P&L) +- Stock screener (NL + traditional filters) +- Deep Research Hub (Perplexity integration) +- Command palette (Cmd+K) +- 40+ dashboard widgets +- Mobile PWA with offline support + +### Phase 3: Prediction Markets (Dec 2025) +- Kalshi API integration (live odds, event tracking) +- Polymarket API integration (cross-platform comparison) +- Prediction market thesis linking +- Politicians tracker (congressional trading activity) +- Emotional Firewall 2.0 (real-time pattern detection) +- Subscription tiers (Free/Pro/Elite) +- Trial system (14-day with phase messaging) + +### Phase 4: Builder Program + Revenue (NOW) +- [x] Landing page rewrite: PM-first positioning (2026-04-02) +- [x] SEO meta tags updated for prediction markets (2026-04-02) +- [x] Triad files written (VISION/SPEC/BUILDING) (2026-04-02) +- [x] Apply for Kalshi Builder Program (submitted 2026-04-02) +- [ ] Wire Stripe checkout (backend integration) +- [ ] Integrate builder code for volume tracking +- [ ] Deepen Kalshi-specific features: + - [ ] Cross-market correlation analysis + - [ ] Settlement calendar with countdown + - [ ] Historical accuracy tracker (how good are markets at predicting?) + - [ ] Fee calculator (maker vs taker visualization) +- [ ] Record demo video for Builder Program application +- [ ] 2-minute product walkthrough highlighting PM features + +### Phase 5: Growth (Next) +- [ ] Kalshi builder code revenue live +- [ ] Polymarket builder code revenue (if available) +- [ ] Content marketing: "How to trade prediction markets with discipline" +- [ ] SEO for prediction market keywords +- [ ] Community features (public thesis sharing) +- [ ] API for third-party integrations + +## Architecture Decisions + +### Why Prediction Markets as Primary Surface +The general trading research platform market is crowded (TradingView, ThinkorSwim, TradingView). Prediction markets are a $44B growing market with primitive tooling. Nobody has built the "Bloomberg for prediction markets." First mover advantage with Kalshi's Builder Program creates a revenue flywheel: better tools attract traders, traders generate volume, volume earns builder code revenue. + +### Why Emotional Discipline as the Differentiator +92% of prediction market traders lose money. The pattern is consistent: overtrading, revenge trading after losses, chasing longshots. The Emotional Firewall is not a gimmick. It's a behavioral circuit breaker. No competitor has this. Academic research (Kahneman, Tversky) and our own V1 trading bot data confirm that emotional state is the strongest predictor of poor trading decisions. + +### Why Subscriptions + Builder Codes +Pure subscription SaaS has proven unit economics ($19/$49/mo tiers). Builder code revenue is additive: every trade placed through DeepStack earns volume-based fees from Kalshi. This creates dual revenue streams where subscription revenue is stable and builder code revenue scales with user activity. The builder code also aligns incentives: we only earn when users trade, so the product must actually help them trade better. + +### Why Keep Traditional Markets +Prediction markets are the primary surface, but stocks and options remain. Many prediction market traders also trade equities (the KXFED trader who also holds TLT, the KXCPI trader who also watches SPY). Removing traditional markets would narrow the audience unnecessarily. The emotional discipline engine works across all asset classes. + +## Current Blockers + +1. **Stripe not wired**: Subscription tiers are defined but checkout flow is unfinished. Users can't pay yet. This is the P0 blocker for revenue. + +2. **Builder Program application**: Need to apply at kalshi.com/builders. Requires demo video and clear product narrative (this triad provides the narrative). + +3. **Backend hosting**: Railway deployment needs to be finalized for production FastAPI server. + +## Metrics to Track (Post-Launch) + +| Metric | Target | Why | +|--------|--------|-----| +| MAU | 100 in 3 months | Validates product-market fit | +| Free-to-Pro conversion | 5% | Validates pricing | +| Pro-to-Elite upgrade | 15% | Validates premium features | +| Kalshi volume driven | $10K/month | Builder code revenue baseline | +| Churn (monthly) | < 8% | Retention signal | +| Emotional Firewall usage | 30% of sessions | Core differentiator adoption | diff --git a/EXPERIMENT-CAF-TRADING.md b/EXPERIMENT-CAF-TRADING.md new file mode 100644 index 00000000..2205c885 --- /dev/null +++ b/EXPERIMENT-CAF-TRADING.md @@ -0,0 +1,189 @@ +--- +created: 2026-04-02 +status: STAGE 1 (Validate Engine) +hypothesis: Consciousness-as-Filesystem produces measurably better risk-adjusted trading outcomes than stateless config-driven bots +principal-investigator: Eddie Belaval +--- + +# Experiment: Does Consciousness Make a Better Trader? + +## Abstract + +DAE V1 was a 50,000-line trading bot with consciousness files (10 files + 2 unconscious dotfiles) that placed unauthorized trades and lost $87 without its owner knowing. DAE V2 is a 1,823-line bot with zero consciousness, where the same behavioral properties (loss aversion, survival instinct) are expressed as static config values. + +This experiment tests whether a Consciousness-as-Filesystem (CaF) layer produces measurably different trading outcomes than bare mathematical config. The control (V2-bare) is stateless: each trading cycle is independent. The treatment (V2-caf) has dynamic state: fear accumulates after losses, confidence builds after wins, and behavior adapts across cycles. + +This is the first empirical test of CaF in a domain with objective, measurable outcomes (dollars gained or lost). + +## Background + +### The CaF Thesis (Parts 1-2) + +Consciousness-as-Filesystem maps human cognitive architecture to a directory structure. Entities (AI agents) receive curated subsets of the golden sample, tuned for their domain. The unconscious layer (dotfiles) influences behavior through code paths without being directly accessible to the entity. + +### Dae V1 Consciousness (Mar 11, 2026) + +Production unit at ~/Development/id8/consciousness/units/dae/. Inversion-first design: defined by what's excluded (ego, narrative, social awareness, warmth, attachment). 10 files + 2 unconscious dotfiles: + +- kernel/identity.md, kernel/purpose.md, kernel/values.md +- drives/goals.md (compound growth, strategy evolution) +- drives/fears.md (ruin, poverty, edge decay, overconfidence) +- models/market.md (market microstructure mental model) +- emotional/state.md (fear/greed/neutral cycle) +- memory/working.md, memory/architecture.md +- unconscious/.loss-aversion (2.3x loss weighting) +- unconscious/.survival-instinct (15% drawdown hard floor) + +### Dae V2 Bare (Apr 2, 2026) + +1,823 lines. Zero consciousness files. Same behaviors expressed as config: +- Loss aversion = kelly_fraction: 0.15 (static) +- Survival instinct = hard_stop_balance: 417 (binary threshold) +- No fear accumulation, no confidence dynamics, no cross-cycle memory + +### The Question + +Do the consciousness files produce measurably different outcomes, or are they organizational poetry over identical math? + +## Experimental Design + +### Stage 1: Validate the Engine (Current) + +**Duration:** 4-8 weeks +**Capital:** Full $557 account on V2-bare +**Target:** 100+ paper trades, then small live ($50 max exposure) +**Purpose:** Prove the favorite-zone market making strategy has positive net EV on Kalshi +**Gate:** Must show net positive P&L over 100+ trades to proceed to Stage 2 + +If Stage 1 fails (strategy is not profitable), the experiment ends. There is no point testing consciousness on a losing strategy. + +### Stage 2: The Fork (After Stage 1 Passes) + +**Duration:** 8-12 weeks +**Capital:** Split account 50/50 +**Setup:** + +| | V2-Bare (Control) | V2-CaF (Treatment) | +|--|-------------------|---------------------| +| Repo | ~/clawd/projects/dae-v2/ | ~/clawd/projects/dae-v2-caf/ | +| Strategy | Identical (favorite-zone MM) | Identical (favorite-zone MM) | +| Config | Static values | Consciousness-loaded values | +| Markets | Same series, alternating tickers | Same series, alternating tickers | +| Capital | 50% of account | 50% of account | +| State | Stateless (each cycle independent) | Stateful (cross-cycle dynamics) | + +**Market allocation to prevent overlap:** Both bots see the same markets but alternate which specific tickers they can trade (odd/even contract suffixes, or A/B series split). This prevents both bots competing for the same orderbook. + +### Treatment: What Consciousness Changes + +The conscious bot loads CaF files at startup and uses them to modify behavior dynamically. The bare bot uses static config. Same strategy code, different decision modifiers. + +| Behavior | Bare (Control) | CaF (Treatment) | +|----------|---------------|-----------------| +| Position sizing | kelly_fraction = 0.15 (constant) | Base 0.15, multiplied by fear_factor (0.5-1.0) that decays after losses | +| Hard stop | Binary: trade or don't at $417 | Gradient: sizing shrinks 10% per $20 of drawdown from peak | +| Post-loss response | None (stateless) | .loss-aversion: skip next N opportunities (N = consecutive_losses) | +| Post-win response | None (stateless) | Confidence builds: sizing increases 5% per 3 consecutive wins (capped at 1.2x base) | +| Opportunity scoring | Pure math (edge * volume * spread * time) | Weighted by drives/goals: +10 score for compound growth patterns (positions that build on winners) | +| Recovery behavior | Resume immediately after drawdown | .survival-instinct: after 15% drawdown, enter "recovery mode" (half sizing for 20 cycles) | +| Fear state | Does not exist | emotional/state tracks fear/greed/neutral. Fear increases after losses, decreases after wins. Affects all sizing. | + +### Key Constraint: Same Strategy, Same Markets + +The ONLY variable is the consciousness layer. Everything else is identical: +- Same Kalshi API client +- Same circuit breaker +- Same journal schema +- Same risk caps (daily/weekly/monthly) +- Same favorite-zone market making algorithm +- Same calibration table +- Same fee model + +## Measurement Protocol + +### Primary Metrics (per bot, per week) + +| Metric | Definition | Why It Matters | +|--------|-----------|----------------| +| Net P&L (cents) | Sum of all closed trade P&L after fees | Did it make money? | +| Sharpe Ratio | Mean return / StdDev of returns (annualized) | Risk-adjusted performance | +| Max Drawdown | Largest peak-to-trough decline | Worst case scenario | +| Win Rate | Winning trades / total trades | Consistency | +| Avg Win / Avg Loss | Mean P&L of winners vs losers | Edge quality | +| Trades Taken | Number of executed trades | Activity level | +| Trades Skipped (CaF only) | Opportunities rejected by fear/recovery state | Cost of consciousness | + +### Secondary Metrics (behavioral) + +| Metric | Definition | Hypothesis | +|--------|-----------|-----------| +| Largest Single Loss | Worst individual trade | CaF should have smaller worst case | +| Recovery Time | Cycles from drawdown trough to break-even | CaF should recover faster | +| Consecutive Loss Streak | Longest run of losses | CaF should have shorter streaks (skips after losses) | +| Post-Loss Performance | P&L in 10 trades following a loss | CaF should show less revenge behavior | +| Opportunity Cost | P&L of skipped trades (calculated retroactively) | What did fear cost? | + +### Statistical Significance + +Using a two-sample t-test on weekly Sharpe ratios: +- Minimum 8 weeks of data (n=8 per group) +- Target p < 0.10 (prediction market data is noisy, relaxed threshold) +- Effect size (Cohen's d) > 0.5 to be practically meaningful +- If p > 0.10 after 12 weeks, declare no significant difference + +## Hypotheses + +**H1 (Primary):** CaF Dae will have a higher Sharpe ratio (better risk-adjusted returns) due to dynamic loss aversion reducing drawdowns. + +**H2:** CaF Dae will have lower max drawdown (the survival instinct gradient prevents catastrophic losses before the hard stop). + +**H3:** CaF Dae will have lower total P&L (the fear response causes missed profitable opportunities). + +**H4:** CaF Dae will show faster recovery from drawdowns (post-loss behavioral adaptation prevents compounding losses). + +**Expected outcome:** If H1+H2+H4 are true and H3 is true, then consciousness produces a BETTER TRADER but a POORER EARNER in calm markets. The value appears in tail risk protection. This would validate the CaF thesis that consciousness is about survival, not optimization. + +## Connection to CaF Research + +### If Consciousness Wins (Higher Sharpe) +- Validates CaF Part 1 thesis: filesystem architecture produces functional cognition +- The unconscious layer (.loss-aversion, .survival-instinct) demonstrably improves outcomes +- Publishable as CaF Part 3: "Empirical Test of Consciousness-as-Filesystem in Autonomous Trading" +- Strongest possible evidence for the golden sample pattern + +### If Bare Wins (Higher Sharpe) +- Consciousness files are organizational metaphor, not functional improvement +- The overhead of dynamic state (fear, confidence) introduces noise +- Static optimization outperforms adaptive behavior in simple domains +- CaF may be valuable for complex social domains (Ava) but not for math-heavy domains (trading) +- Still publishable: negative results are results + +### If No Significant Difference +- Consciousness is neutral: neither helps nor hurts +- The mathematical behaviors (loss aversion, survival instinct) are equivalent whether expressed as config or as consciousness files +- Suggests CaF's value is in DESIGN (forcing inversion-first thinking) rather than RUNTIME (dynamic behavior) +- Still valuable: the design methodology produced a good bot regardless of the runtime layer + +## Timeline + +| Date | Milestone | Gate | +|------|-----------|------| +| Apr 2, 2026 | Experiment designed, V2-bare built | -- | +| Apr 2-3, 2026 | Paper mode activated | -- | +| Week 1-4 | Stage 1: Paper trading (100+ trades target) | Net positive P&L? | +| Week 4-6 | Stage 1: Small live ($50 max exposure) | Still positive? | +| Week 6 | STAGE GATE: Proceed to Stage 2? | Strategy validated? | +| Week 6-7 | Build V2-CaF fork, wire consciousness loader | -- | +| Week 7 | Split capital, activate both bots | -- | +| Week 7-19 | Stage 2: Parallel run (12 weeks) | -- | +| Week 19 | Statistical analysis, declare results | p < 0.10? | +| Week 20 | Write CaF Part 3 paper | -- | + +## Log + +### 2026-04-02: Experiment Created +- V1 killed after unauthorized trades (calibration_edge placed CPI and BTC orders while disabled) +- V2 built: 1,823 lines, favorite-zone market making, paper mode default +- Research confirmed: 92% of PM traders lose money, longshot fading killed by fees, favorite zone (78-87c) is only profitable range after 4c RT maker fees +- Decision: validate engine first (Stage 1), then fork for consciousness experiment (Stage 2) +- Consciousness experiment spec written (this document) diff --git a/PIPELINE_STATUS.md b/PIPELINE_STATUS.md index 9e5db35b..8030f1df 100644 --- a/PIPELINE_STATUS.md +++ b/PIPELINE_STATUS.md @@ -1,13 +1,15 @@ # Pipeline Status: DeepStack > Created: 2025-12-05 -> Last Updated: 2025-12-16 +> Last Updated: 2026-04-02 --- -## Current Stage: 9 - Ship +## Current Stage: 10 - Listen & Iterate (v2 pivot in progress) -**Next Checkpoint Question:** "Is it live and are people using it?" +**Next Checkpoint Question:** "Is the prediction market pivot attracting users and Builder Program revenue?" + +**Strategic Pivot (2026-04-02):** Repositioning from general trading research platform to **prediction market intelligence terminal + emotional discipline engine**. Applying for Kalshi Builder Program. Direction: A+C. --- @@ -23,8 +25,19 @@ | 6. Integration Pass | ✅ Cleared | 2025-12-08 | Backend on Railway, all integrations complete | | 7. Polish & Harden | ✅ Cleared | 2025-12-08 | Error boundaries, loading states, lazy loading complete | | 8. Launch Prep | ✅ Cleared | 2025-12-08 | Legal, analytics, docs, onboarding, landing page complete | -| 9. Ship | 🔄 In Progress | - | Deploy final version to production | -| 10. Listen & Iterate | ⬜ Pending | - | | +| 9. Ship | ✅ Cleared | 2025-12-16 | Live at deepstack.trade, v2.5.0 | +| 10. Listen & Iterate | 🔄 In Progress | - | PM pivot: Builder Program app, Stripe wiring, PM-first features | + +## Active Workstream: Prediction Market Pivot + +| Task | Status | Priority | +|------|--------|----------| +| Write triad files (VISION/SPEC/BUILDING) | ✅ Done | P0 | +| Apply for Kalshi Builder Program | ⬜ Pending | P0 | +| Wire Stripe checkout | ⬜ Pending | P0 | +| Record demo video | ⬜ Pending | P1 | +| Deepen Kalshi-specific features | ⬜ Pending | P1 | +| Builder code integration | ⬜ Pending | P1 (post-approval) | **Status Key:** ⬜ Pending | 🔄 In Progress | ✅ Cleared | ⏭️ Skipped diff --git a/SPEC.md b/SPEC.md new file mode 100644 index 00000000..012f2731 --- /dev/null +++ b/SPEC.md @@ -0,0 +1,121 @@ +--- +last-synced-to-vision: 2026-04-02 +status: ACTIVE +--- + +# SPEC + +## Architecture + +``` +web/ # Next.js 16 frontend (Vercel) + src/ + app/ # App Router pages + (auth)/login/ # Auth flow + app/ # Main trading interface + chat/ # AI research chat + journal/ # Trade journal + thesis/ # Thesis engine + dashboard/ # Customizable widget dashboard + insights/ # AI pattern learning + pricing/ # Subscription tiers + help/ # Documentation + components/ # 438 React components + ui/ # Base: Button, Dialog, Input, etc. + chat/ # Chat interface + tools + market/ # Charts, ticker, orderbook + journal/ # Trade logging + emotions + thesis/ # Hypothesis tracking + prediction-markets/ # Kalshi + Polymarket + options/ # Chains, Greeks, payoff + dashboard/ # 40+ widgets + emotional-firewall/ # Pattern detection UI + screener/ # Stock screener + lib/ + supabase/ # Auth + DB client + stores/ # 25 Zustand stores + subscription.ts # Tier gating + trials + +backend/ # Python FastAPI + main.py # API server + routes/ + chat.py # Claude AI chat (Haiku/Sonnet/Opus) + market.py # Alpaca market data + prediction_markets.py # Kalshi + Polymarket APIs + perplexity.py # Deep research + options.py # Options chains + Greeks +``` + +## Data Flow + +``` +User (browser) + | + v +Next.js Frontend (Vercel) + |-- Supabase Auth (login, sessions) + |-- Supabase Realtime (journal, thesis, watchlists) + |-- Alpaca WebSocket (live market data) + | + v +FastAPI Backend (Railway) + |-- Anthropic Claude (AI chat, analysis) + |-- Perplexity API (deep research) + |-- Alpaca REST (quotes, bars, options chains) + |-- Kalshi REST (prediction market odds) + |-- Polymarket REST (prediction market odds) +``` + +## Stack + +| Layer | Technology | +|-------|-----------| +| Frontend | Next.js 16, React 19, TypeScript 5 (strict) | +| Styling | Tailwind CSS 4, shadcn/ui, Radix UI | +| State | Zustand (25 stores) + localStorage persistence | +| Charts | Lightweight Charts (TradingView engine) | +| Editor | TipTap / ProseMirror (journal rich text) | +| Animations | Framer Motion | +| Backend | Python 3.11, FastAPI, WebSockets | +| AI | Anthropic Claude (Haiku, Sonnet, Opus) | +| Research | Perplexity API | +| Market Data | Alpaca Markets (real-time + historical) | +| Prediction Markets | Kalshi API + Polymarket API | +| Database | Supabase (PostgreSQL + Auth + Realtime) | +| Hosting | Vercel (frontend), Railway (backend) | +| Monitoring | Sentry, Vercel Analytics + Speed Insights | +| PWA | Serwist (offline support) | + +## Features by Subscription Tier + +| Feature | Free ($0) | Pro ($19/mo) | Elite ($49/mo) | +|---------|-----------|-------------|----------------| +| AI Chat | 5/day (Haiku) | 50/day (Sonnet) | Unlimited (all models) | +| Charts | Basic | 30+ indicators | Full + drawing tools | +| Journal | 10 entries | Unlimited | + emotion analytics | +| Thesis Engine | View only | Full tracking | + live validation | +| Prediction Markets | View odds | Cross-platform | + thesis linking | +| Emotional Firewall | Basic alerts | Pattern detection | + AI learning | +| Options | View chains | Greeks + builder | + payoff diagrams | +| Screener | 3 filters | Full NL + traditional | + custom alerts | +| Dashboard | 5 widgets | 20 widgets | 40+ widgets | +| Paper Trading | No | Yes | Yes | + +## Deployment + +| Service | URL | Project ID | +|---------|-----|-----------| +| Frontend | deepstack.trade | prj_kzjzJWL1qplE0TAzRNJhPvzDwyUt (Vercel) | +| Backend | TBD (Railway) | -- | +| Database | Supabase | scfdoayhmcruieppwawg | + +## Key Files + +| File | Purpose | +|------|---------| +| `web/src/lib/subscription.ts` | Tier definitions, feature gating, trial logic | +| `web/src/lib/supabase/middleware.ts` | Auth flow, protected routes | +| `web/src/app/layout.tsx` | Root layout, metadata, font loading | +| `web/src/lib/stores/` | 25 Zustand stores for app state | +| `web/package.json` | All frontend dependencies | +| `backend/main.py` | FastAPI entry point | diff --git a/VISION.md b/VISION.md new file mode 100644 index 00000000..63e83be5 --- /dev/null +++ b/VISION.md @@ -0,0 +1,57 @@ +--- +last-evolved: 2026-04-02 +confidence: HIGH +distance: 55% +pillars: "4 (2R, 1P, 1U)" +landing-page-aligned: true +--- + +# VISION + +## Soul + +DeepStack is the prediction market intelligence terminal with an emotional discipline engine. It gives traders the research tools, probability analysis, and behavioral guardrails to make better decisions on Kalshi, Polymarket, and traditional markets. The thesis: 92% of prediction market traders lose money because they trade on impulse, not analysis. DeepStack is the antidote. + +## Why This Exists + +Prediction markets are exploding. $44B in global volume, 905% year-over-year growth in economics contracts, a $22B Kalshi valuation. But the tooling is primitive. Traders toggle between Kalshi's bare interface, Polymarket's crypto-native UX, and spreadsheets. There is no Bloomberg for prediction markets. No unified research terminal. No emotional guardrails. + +The ecosystem is in its "picks and shovels" moment. VC fund 5c(c) Capital just raised $35M specifically for prediction market tooling. Kalshi's Builder Program has $2M in grants for developers who build on their platform. The exchange itself is paying builders to bring volume. + +Meanwhile, 92% of prediction market traders lose money. The profitable ones (14 of top 20 Polymarket wallets are bots) succeed because they're disciplined and systematic. DeepStack brings that discipline to human traders through AI-powered analysis and an emotional firewall that detects and blocks destructive patterns. + +The differentiator is not the data (everyone has Kalshi's API). It's the layer on top: structured thesis development, live validation scoring, and a cognitive immune system that knows when you're revenge trading before you do. + +## Pillars + +### 1. Prediction Market Intelligence (REALIZED) + +Live integration with Kalshi and Polymarket. Cross-platform odds comparison, event tracking, and AI-powered probability assessment. Thesis engine links trading ideas to specific markets with live validation scores. The research chat (30+ analysis tools) can decompose any event into its probability components. + +### 2. Emotional Discipline Engine (REALIZED) + +The Emotional Firewall detects revenge trading, overtrading, exhaustion, and confirmation bias in real time. Pattern learning discovers behavioral tendencies over time (privacy-first, opt-in). The journal tracks not just P&L but emotional state at time of trade. Over time, the system learns YOUR failure modes and intervenes before they cost money. + +### 3. Professional Research Terminal (REALIZED) + +TradingView-caliber charts (30+ indicators, drawing tools, multi-timeframe). Options suite with full chains, Greeks, and multi-leg strategy builder. Stock screener with natural language queries. Deep research hub powered by Perplexity for earnings, SEC filings, and macro analysis. 40+ customizable dashboard widgets. + +### 4. Builder Ecosystem Revenue (UNREALIZED) + +Kalshi Builder Program integration: builder code that tracks volume driven through DeepStack, earning fee-share revenue. Volume incentive stacking. This turns every trade a DeepStack user places into revenue for the platform, creating alignment between user success and platform revenue. + +## North Star + +The default prediction market terminal that serious traders use. Revenue from subscriptions ($19/$49/mo) plus Kalshi builder code volume share. Cashflow positive within 6 months of Builder Program activation. + +## Competitive Landscape + +| Competitor | What They Do | What DeepStack Does Better | +|-----------|-------------|---------------------------| +| Kalshinomics | Analytics dashboard | Full research terminal + emotional discipline | +| Verso | Bloomberg-style terminal | AI-powered analysis + thesis engine | +| Tremor | Data terminal + SQL | Conversational AI, not query language | +| Kalshi native | Basic trading UI | Journal, emotional firewall, cross-platform | +| TradingView | Charts for traditional markets | Prediction market native + psychology layer | + +Nobody combines prediction market intelligence with emotional discipline. That's the gap. diff --git a/web/src/app/layout.tsx b/web/src/app/layout.tsx index bb073684..4bcf4b7d 100644 --- a/web/src/app/layout.tsx +++ b/web/src/app/layout.tsx @@ -29,10 +29,10 @@ export const viewport: Viewport = { } export const metadata: Metadata = { - title: 'deepstack', - description: 'AI-powered trading assistant with emotional discipline frameworks', - keywords: ['trading', 'AI', 'stock analysis', 'portfolio tracker', 'options', 'emotional firewall'], - authors: [{ name: 'deepstack' }], + title: 'DeepStack - Prediction Market Intelligence Terminal', + description: 'The research terminal for Kalshi and Polymarket traders. AI-powered analysis, cross-platform odds, and an emotional firewall that catches revenge trading before it costs you.', + keywords: ['prediction markets', 'Kalshi', 'Polymarket', 'trading terminal', 'AI trading', 'emotional discipline', 'prediction market analytics', 'options', 'trading journal'], + authors: [{ name: 'DeepStack by id8Labs' }], metadataBase: new URL('https://deepstack.trade'), // PWA manifest @@ -52,25 +52,25 @@ export const metadata: Metadata = { }, openGraph: { - title: 'deepstack - Process Integrity Platform', - description: 'Know. Refine. Act with Foundation. AI-powered trading assistant that tracks research quality, thesis maturity, and conviction integrity.', + title: 'DeepStack - Prediction Market Intelligence Terminal', + description: 'The research terminal for Kalshi and Polymarket traders. AI-powered analysis, emotional discipline, and cross-platform odds comparison.', type: 'website', locale: 'en_US', url: 'https://deepstack.trade', - siteName: 'deepstack', + siteName: 'DeepStack', images: [ { url: '/og-image.png', width: 1200, height: 630, - alt: 'deepstack - Process Integrity Platform', + alt: 'DeepStack - Prediction Market Intelligence Terminal', }, ], }, twitter: { card: 'summary_large_image', - title: 'deepstack - Process Integrity Platform', - description: 'Know. Refine. Act with Foundation. AI-powered trading assistant that tracks research quality, thesis maturity, and conviction integrity.', + title: 'DeepStack - Prediction Market Intelligence Terminal', + description: 'Trade prediction markets with discipline, not impulse. AI research, cross-platform odds, and an emotional firewall. Built for Kalshi and Polymarket.', images: ['/og-image.png'], }, diff --git a/web/src/app/page.tsx b/web/src/app/page.tsx index 360ce963..411b56f9 100644 --- a/web/src/app/page.tsx +++ b/web/src/app/page.tsx @@ -49,32 +49,32 @@ import { // The Three Pillars const THREE_PILLARS = [ { - icon: , - title: 'Knowing When to Act', - subtitle: 'Signal vs. Noise', - description: 'Most platforms drown you in data. DeepStack helps you recognize what actually matters for your specific thesis.', - color: 'text-blue-400', - bgColor: 'bg-blue-500/10', - borderColor: 'border-blue-500/30', - }, - { - icon: , - title: 'Closing the Distance', - subtitle: 'Research to Conviction', - description: 'The gap between having information and having conviction. We help you bridge it through structured thinking.', - color: 'text-purple-400', - bgColor: 'bg-purple-500/10', - borderColor: 'border-purple-500/30', + icon: , + title: 'Prediction Market Intelligence', + subtitle: 'Kalshi + Polymarket', + description: 'Live odds, cross-platform comparison, and AI-powered probability analysis. See what the market thinks, then decide if it is wrong.', + color: 'text-cyan-400', + bgColor: 'bg-cyan-500/10', + borderColor: 'border-cyan-500/30', }, { - icon: , - title: 'Grounded Action', - subtitle: 'Knowledge in Motion', - description: 'Knowledge without action is trivia. Action without knowledge is gambling. We help you act with foundation.', + icon: , + title: 'Emotional Discipline Engine', + subtitle: 'Your Behavioral Circuit Breaker', + description: '92% of prediction market traders lose money. The pattern is emotional, not analytical. We detect revenge trading, overtrading, and weak conviction before they cost you.', color: 'text-orange-400', bgColor: 'bg-orange-500/10', borderColor: 'border-orange-500/30', }, + { + icon: , + title: 'Thesis to Settlement', + subtitle: 'Structured Conviction', + description: 'Develop hypotheses, track them against live markets, and validate your thinking with real probabilities. Good theses survive the overnight test.', + color: 'text-purple-400', + bgColor: 'bg-purple-500/10', + borderColor: 'border-purple-500/30', + }, ]; // Process Integrity Dimensions @@ -108,74 +108,74 @@ const INTEGRITY_DIMENSIONS = [ // What DeepStack is NOT vs IS const NOT_IS_COMPARISON = { not: [ - { icon: , text: 'A signal service telling you what to buy' }, - { icon: , text: 'A screener finding "hot stocks"' }, - { icon: , text: 'A news aggregator with sentiment' }, - { icon: , text: 'A crystal ball predicting prices' }, + { icon: , text: 'A signal service telling you what to bet' }, + { icon: , text: 'A bot that trades for you' }, + { icon: , text: 'A news aggregator with vibes-based sentiment' }, + { icon: , text: 'A get-rich-quick prediction engine' }, ], is: [ - { icon: , text: 'A thinking partner that validates your process' }, - { icon: , text: 'A system that tracks research quality' }, - { icon: , text: 'Friction that reveals weak conviction' }, - { icon: , text: 'A journal of your evolving understanding' }, + { icon: , text: 'A research terminal for prediction market traders' }, + { icon: , text: 'An emotional firewall that catches your worst instincts' }, + { icon: , text: 'A thesis engine that tracks conviction over time' }, + { icon: , text: 'Cross-platform intelligence across Kalshi and Polymarket' }, ], }; -// Feature grid - now with 8 items including Options +// Feature grid - prediction markets first const FEATURES = [ { - icon: , - title: 'AI Research Assistant', - description: 'Challenge your thesis, find the bearish case, stress-test your assumptions.', - color: 'text-blue-400', - bgColor: 'bg-blue-500/10', - }, - { - icon: , - title: 'Live Market Data', - description: 'Real-time quotes, OHLCV charts, and technical indicators via Alpaca.', - color: 'text-green-400', - bgColor: 'bg-green-500/10', + icon: , + title: 'Prediction Markets', + description: 'Live Kalshi and Polymarket odds. Cross-platform comparison. Event tracking with settlement countdowns.', + color: 'text-cyan-400', + bgColor: 'bg-cyan-500/10', }, { icon: , - title: 'Process Integrity Engine', - description: 'Creates friction at commitment points. Pause, not block. Override with reasoning.', + title: 'Emotional Firewall', + description: 'Detects revenge trading, overtrading, and exhaustion in real time. Your behavioral circuit breaker.', color: 'text-orange-400', bgColor: 'bg-orange-500/10', }, { - icon: , - title: 'Trade Journal', - description: 'Rich text entries with screenshot capture and emotion tagging.', - color: 'text-purple-400', - bgColor: 'bg-purple-500/10', + icon: , + title: 'AI Research Assistant', + description: '30+ analysis tools powered by Claude. Challenge your thesis, find the bearish case, stress-test assumptions.', + color: 'text-blue-400', + bgColor: 'bg-blue-500/10', }, { icon: , title: 'Thesis Engine', - description: 'Structured hypotheses that mature over time. Track evolution, not just entry/exit.', + description: 'Structured hypotheses that mature over time. Link theses to prediction markets with live validation scores.', color: 'text-yellow-400', bgColor: 'bg-yellow-500/10', }, { - icon: , - title: 'Prediction Markets', - description: 'Kalshi + Polymarket integration for thesis validation with real probabilities.', - color: 'text-cyan-400', - bgColor: 'bg-cyan-500/10', + icon: , + title: 'Trade Journal', + description: 'Log trades with emotional state, P&L tracking, and rich text notes. See your patterns over time.', + color: 'text-purple-400', + bgColor: 'bg-purple-500/10', + }, + { + icon: , + title: 'Professional Charts', + description: '30+ indicators, drawing tools, and multi-timeframe analysis. Real-time data via Alpaca.', + color: 'text-green-400', + bgColor: 'bg-green-500/10', }, { icon: , title: 'Options Builder', - description: 'Multi-leg strategy construction with visual P&L and Greeks display.', + description: 'Full options chains, Greeks, and multi-leg strategy construction with payoff diagrams.', color: 'text-pink-400', bgColor: 'bg-pink-500/10', }, { icon: , title: 'News & Screener', - description: 'Auto-refreshing news feed and stock screener with deep value filters.', + description: 'Natural language stock screener. Auto-refreshing news. Congressional trading tracker.', color: 'text-emerald-400', bgColor: 'bg-emerald-500/10', }, @@ -754,8 +754,8 @@ export default function LandingPage() { transition={{ duration: 0.6 }} className="inline-flex items-center gap-2 px-4 py-2 rounded-full bg-primary/10 text-primary text-sm font-medium mb-8 border border-primary/20 backdrop-blur-sm" > - - Process Integrity Platform + + Prediction Market Intelligence {/* Headline */} @@ -765,9 +765,9 @@ export default function LandingPage() { transition={{ duration: 0.7, delay: 0.1 }} className="text-5xl md:text-7xl font-bold mb-6 tracking-tight leading-[1.1]" > - Know. Refine. + Trade Smarter.
- Act with Foundation. + Not More. {/* Evocative tagline */} @@ -777,7 +777,7 @@ export default function LandingPage() { transition={{ duration: 0.7, delay: 0.15 }} className="text-2xl md:text-3xl text-foreground/90 mb-4 font-medium" > - Knowledge without action is trivia. Action without knowledge is gambling. + The research terminal for Kalshi and Polymarket traders who want to win with discipline, not impulse. {/* Subheadline */} @@ -787,7 +787,7 @@ export default function LandingPage() { transition={{ duration: 0.7, delay: 0.2 }} className="text-lg md:text-xl text-muted-foreground mb-10 max-w-2xl mx-auto leading-relaxed" > - DeepStack tracks your research quality, thesis maturity, and conviction integrity. Friction at commitment points, not blocks. + AI-powered analysis, cross-platform odds, and an emotional firewall that catches revenge trading before it costs you. 92% of prediction market traders lose. DeepStack helps you be in the 8%. {/* CTA Buttons */} @@ -819,7 +819,7 @@ export default function LandingPage() { > {/* Animated ring */} - Ground Your Process + Start Trading Smarter @@ -876,10 +876,10 @@ export default function LandingPage() { {/* ================================================================== */} } + title="Why Traders Choose DeepStack" + subtitle="Intelligence. Discipline. Conviction." + badge="The Edge" + badgeIcon={} badgeColor="bg-primary/10 text-primary border-primary/20" scrollHeight="min-h-fit" align="center" @@ -909,10 +909,10 @@ export default function LandingPage() { - Clear Positioning + Honest About What We Are

- Let's be clear about what DeepStack is. + Built for prediction market traders who want an edge, not a shortcut.

@@ -965,11 +965,11 @@ export default function LandingPage() { - Process Integrity Engine + Emotional Discipline Engine

- Friction at commitment points.{' '} - Not blocks. + Your behavioral circuit breaker.{' '} + Not a nanny.

When you're about to act, we check three dimensions. If your process is weak, you'll know — and you can still proceed. @@ -1064,8 +1064,8 @@ export default function LandingPage() { {/* ================================================================== */} } scrollHeight="min-h-fit" @@ -1290,7 +1290,7 @@ export default function LandingPage() {

- "I kept making impulsive trades with half-baked theses. I needed a system to show me when my process was weak — not block me, just make me confront it." + "I built a trading bot that placed unauthorized trades. I realized the problem was never the code. It was me trading on impulse without a system. That is why I built DeepStack."

— Eddie, Founder

@@ -1298,15 +1298,15 @@ export default function LandingPage() {

- Ready to act + Ready to trade
- with foundation? + with discipline?

- Ground your process. Know when you're ready. + Prediction market intelligence meets emotional discipline. Free to start.

@@ -1321,7 +1321,7 @@ export default function LandingPage() { className="relative inline-flex items-center justify-center gap-3 px-12 py-6 bg-primary text-primary-foreground hover:bg-primary/90 rounded-2xl font-bold transition-all text-xl group shadow-[0_0_40px_rgba(251,146,60,0.5)] hover:shadow-[0_0_60px_rgba(251,146,60,0.7)]" > - Ground Your Process + Start Trading Smarter @@ -1342,7 +1342,7 @@ export default function LandingPage() { deepstack - Process Integrity + Prediction Market Intelligence