API Reference

curl -X POST https://agent-toolbelt-production.up.railway.app/api/tools/stock-thesis \
  -H "Authorization: Bearer atb_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"ticker": "NVDA"}'

{
  "ticker": "NVDA",           // required — US stock ticker
  "timeHorizon": "3-5 years"  // optional — "1-2 years" | "3-5 years" | "5+ years"
}

{
  "ticker": "NVDA",
  "verdict": "bullish",           // "bullish" | "neutral" | "bearish"
  "oneLiner": "Nvidia owns the essential infrastructure for the AI revolution with a defensible software moat, but the valuation demands flawless execution.",
  "thesis": "Nvidia is in a rare position: it supplies the picks and shovels for an industry-wide megatrend. The company has grown revenue at a staggering 10,005% CAGR over three years... CUDA, its software platform, locks customers into its ecosystem. This network effect is Nvidia's moat.",
  "keyStrengths": [
    "Dominant ~80%+ data center GPU market share",
    "CUDA moat creates switching costs and customer lock-in",
    "42 buy / 5 hold / 1 sell analyst consensus"
  ],
  "keyRisks": [
    "Valuation pricing in near-perfection at 36x earnings",
    "AMD and custom silicon (Google TPU, AWS Trainium) eroding share",
    "Hyperscaler capex slowdown would compress data center revenue"
  ],
  "valuationRead": "Premium but not irrational. 36.9x P/E justified by AI tailwinds — fair value hinges on sustained data center spending through 2029.",
  "insiderRead": "Mixed signals: two executives bought ~47k shares each on March 9 (positive alignment), offset by routine selling from others — likely tax diversification rather than conviction-based.",
  "watchFor": "Data center revenue growth rate and gross margin in the next earnings report. If deceleration goes below 30% YoY, the boom is maturing faster than expected.",
  "dataSnapshot": {
    "marketCapBillions": 2780.4,
    "currentPrice": 113.60,
    "peRatio": 36.9,
    "analystConsensus": { "buy": 42, "hold": 5, "sell": 1 }
  },
  "dataSources": {
    "fetchedAt": "2026-03-22T17:30:00.000Z",
    "polygon": { "success": true },
    "finnhub": { "success": true },
    "fmp": { "success": true }
  },
  "generatedAt": "2026-03-22T17:30:05.123Z"
}

{
  "ticker": "NVDA"  // required
}

{
  "ticker": "NVDA",
  "verdict": "strong_compounder",  // "strong_compounder" | "consistent" | "mixed" | "volatile" | "deteriorating"
  "oneLiner": "Nvidia has delivered 9 consecutive EPS beats with accelerating revenue — a near-perfect earnings track record driven by AI infrastructure demand.",
  "beatRate": "9/10 quarters beat (90%)",
  "revenueTrend": "accelerating",  // "accelerating" | "steady" | "decelerating" | "declining"
  "revenueRead": "Revenue has grown from $4.3B to $35.1B over 8 quarters — a trajectory driven almost entirely by data center GPU demand. No signs of deceleration yet.",
  "epsRead": "Consistently beating estimates by wide margins (avg +18%). Management has been conservative with guidance, setting up repeated upside surprises.",
  "lastQuarterSummary": "Q4 FY2025: EPS $0.89 vs $0.84 estimate (+6% beat). Revenue $35.1B vs $33.5B estimate. Data center up 93% YoY.",
  "longTermRead": "For a long-term holder, this is as clean an earnings history as you'll find in mega-cap tech. The risk is deceleration from a very high base, not deterioration.",
  "watchForNext": "Revenue growth rate — specifically whether data center can sustain 50%+ YoY. A miss or weak guide would be the first crack in the thesis.",
  "rawData": {
    "quartersAnalyzed": 10,
    "beatsTotal": 9,
    "upcomingEarnings": { "date": "2026-05-28", "epsEstimate": 0.93 }
  },
  "dataSources": {
    "fetchedAt": "2026-03-22T17:30:00.000Z",
    "fmp": { "success": true },
    "finnhub": { "success": true }
  },
  "generatedAt": "2026-03-22T17:30:04.891Z"
}

{
  "ticker": "NVDA"  // required
}

{
  "ticker": "NVDA",
  "signal": "neutral",  // "strong_buy" | "buy" | "neutral" | "sell" | "strong_sell"
  "confidence": "medium",  // "high" | "medium" | "low"
  "oneLiner": "Two small open-market purchases offset by heavy routine selling — net signal is neutral.",
  "interpretation": "Two executives made open-market purchases of ~47k shares each in March — a meaningful show of conviction at current prices. However, these are offset by significantly larger selling activity from other insiders, most of which appears routine (tax withholding on vested RSUs, planned 10b5-1 sales).",
  "buyingPressure": "2 open-market purchases totaling ~94k shares (~$10.7M) from two senior executives in March 2026.",
  "sellingPressure": "Multiple sales totaling ~320k shares, predominantly from 10b5-1 plans and RSU tax withholding — routine, not conviction-based.",
  "notableTrades": [
    {
      "who": "Jensen Huang (CEO)",
      "action": "Open-market purchase: +47,210 shares @ $113.40 ($5.4M)",
      "significance": "CEO buying in the open market is a strong alignment signal — not a routine exercise."
    }
  ],
  "verdict": "The purchases are encouraging but insufficient to override the broader selling pattern. For a long-term holder, this is a mild positive — executives aren't fleeing, but the net insider flow is not a strong buy signal.",
  "rawData": {
    "transactionsAnalyzed": 20,
    "openMarketPurchases": 2,
    "openMarketSales": 5,
    "routineTransactions": 13,
    "netSharesPurchased": -226000
  },
  "dataSources": {
    "fetchedAt": "2026-03-22T17:30:00.000Z",
    "finnhub": { "success": true }
  },
  "generatedAt": "2026-03-22T17:30:03.441Z"
}

{
  "ticker": "NVDA"  // required
}

{
  "ticker": "NVDA",
  "companyName": "NVIDIA Corporation",
  "verdict": "expensive",  // "very_cheap" | "cheap" | "fair" | "expensive" | "very_expensive"
  "oneLiner": "NVDA trades at a significant premium to the market — justified only if AI infrastructure spending sustains 50%+ growth through 2027.",
  "valuationRead": "At 36.9x earnings and 28x sales, NVDA is priced for near-perfection. The FCF yield of 2.8% is thin for a cyclical business. ROE of 127% is exceptional and partially justifies the premium, but the multiple leaves no margin of safety.",
  "buyZone": "Below $85 (25x forward earnings) would represent fair value with a margin of safety. At current levels, you're paying for 3+ years of growth in advance.",
  "bottomLine": "Hold if you own it — the business is exceptional. Initiate a position only on a meaningful pullback. At $113, you're not getting a bargain.",
  "metrics": {
    "peRatio": 36.9,
    "psRatio": 28.1,
    "pbRatio": 42.3,
    "evEbitda": 31.4,
    "fcfYield": 2.8,
    "roe": 127.4,
    "netMargin": 55.0,
    "debtToEquity": 0.41
  },
  "dataSources": {
    "fetchedAt": "2026-03-22T17:30:00.000Z",
    "fmp": { "success": true },
    "finnhub": { "success": true },
    "polygon": { "success": true }
  },
  "generatedAt": "2026-03-22T17:30:04.210Z"
}

{
  "ticker": "NVDA"  // required
}

{
  "ticker": "NVDA",
  "bullCase": [
    {
      "argument": "CUDA moat is effectively unbreakable",
      "evidence": "4M+ CUDA developers, 20 years of ecosystem investment, switching cost is years of retraining and retooling. AMD's ROCm has existed for a decade with minimal share gain."
    },
    {
      "argument": "Sovereign AI wave is a new, undiscounted demand driver",
      "evidence": "Countries building national AI infrastructure (UAE, Saudi Arabia, France, India) represent demand not captured in hyperscaler growth estimates."
    },
    {
      "argument": "10,005% revenue CAGR over 3 years — with no deceleration yet",
      "evidence": "Q4 FY2025 data center revenue up 93% YoY. If this sustains even 40% growth, the stock is cheap at current multiples by 2027."
    }
  ],
  "bearCase": [
    {
      "argument": "Hyperscaler capex is inherently cyclical",
      "evidence": "Cloud spending cycles historically compress 30-50%. If AWS, Azure, and Google pause GPU orders for 2 quarters, NVDA revenue could fall 25-40%."
    },
    {
      "argument": "Custom silicon is a real long-term threat",
      "evidence": "Google's TPUs, Amazon's Trainium, and Meta's MTIA are purpose-built and improving rapidly. By 2028, hyperscalers could run 30-40% of workloads on custom chips."
    },
    {
      "argument": "36x earnings gives you zero margin of safety",
      "evidence": "Any earnings miss, guidance cut, or macro shock sends this stock down 20-30%. The multiple embeds perfection across multiple years of execution."
    }
  ],
  "verdict": "slight_bull",  // "bull_wins" | "slight_bull" | "too_close" | "slight_bear" | "bear_wins"
  "verdictRationale": "The bull case wins on business quality and momentum. The CUDA moat is real and the revenue trajectory is exceptional. The bear case is right on valuation risk — this is not a stock for the risk-averse. Net: the business justifies a position, the price demands patience.",
  "keyDebate": "Will hyperscaler AI capex sustain 40%+ growth through 2027, or is this a classic infrastructure overbuild that corrects hard in 2026?",
  "dataSources": {
    "fetchedAt": "2026-03-22T17:30:00.000Z",
    "polygon": { "success": true },
    "fmp": { "success": true },
    "finnhub": { "success": true }
  },
  "generatedAt": "2026-03-22T17:30:05.780Z"
}