AI Hedge Fund Research Agent (US Stocks)

Brief

Get a clear Buy / Hold / Sell view for one US stock ticker in a single run, with confidence score, thesis, catalysts, risks, and source links.

Who is this for

• Investors and traders who want fast, structured stock research
• Product teams building stock insight features into apps
• Automation users who need one reliable decision packet per ticker

What it gives you

• One recommendation packet per run
• Confidence score from 0 to 100
• A short thesis explaining the call
• Upside catalysts and downside risks
• Source links so you can verify the evidence

Input

• ticker (required): US stock symbol (example: AAPL, TSLA, JPM)
• horizon (required): scalping | day_trade | swing | long_term
• maxDrawdownPct (optional): 0-100
• targetVolatilityPct (optional): 0-100
• maxPositionPct (optional): 0-100

Output

Each successful run returns exactly one item with:

• recommendation: buy | hold | sell
• confidencePct: integer 0-100
• thesis: plain-language summary
• catalysts: list of potential upside drivers
• risks: list of key risks/invalidation points
• sources: list of links used by the run

Data Sources

• Public market price data
• Public company fundamentals snapshot
• Recent market news headlines
• Reddit discussion signals

Input example

{
  "ticker": "AAPL",
  "horizon": "swing"
}

Output example

{
  "ticker": "AAPL",
  "horizon": "swing",
  "recommendation": "hold",
  "confidencePct": 58,
  "thesis": "AAPL shows a mixed swing setup with positive recent momentum but stretched conditions.",
  "catalysts": [
    "Recent 20-day momentum is positive.",
    "News flow is supportive."
  ],
  "risks": [
    "Short-term overbought conditions.",
    "Signal quality can vary when public endpoints are rate-limited."
  ],
  "sources": [
    {
      "name": "Price data",
      "url": "https://query1.finance.yahoo.com/v8/finance/chart/AAPL?interval=1d&range=6mo"
    },
    {
      "name": "News feed",
      "url": "https://news.google.com/rss/search?q=AAPL%20stock&hl=en-US&gl=US&ceid=US:en"
    }
  ]
}

FAQ

Q: Does this place trades for me?
A: No. It only provides research output.

Q: Do I need paid market data subscriptions?
A: No. It is designed to run on public/free data sources.

Q: Why are some fields sometimes empty?
A: Some public endpoints can throttle or block requests. The Actor still returns a valid packet using available evidence.

Q: How many items does one run produce?
A: Exactly one decision packet per successful run.

Q: Is this financial advice?
A: No. Use this as research support, not as a guarantee of outcomes.

Release notes

v0.4

• Improved reliability when market or social data endpoints are temporarily blocked
• Added fallback fundamentals coverage so runs return richer company context more consistently
• Improved recommendation generation stability with stricter output validation and retry handling

v0.3

• Updated documentation for end users with simpler wording
• Added clearer examples of input and output payloads
• Improved publication-ready listing description

v0.2

• Added monetization-ready pricing events
• Improved result consistency and validation of output fields
• Improved explanation quality with structured thesis/catalysts/risks format

v0.1

• Initial release with single-ticker decision packet
• Horizon-aware recommendations (scalping, day_trade, swing, long_term)
• Source-linked evidence output for transparency