Multi-agent trading research

Watch four analysts argue your next trade.

Four LLM analysts debate any ticker, then hand you a BUY, HOLD, or SELL with the full transcript.

run / NVDA / 2026-06-06 / deepsample
BUY0.71

Improving fundamentals outweigh near-term sentiment risk. Entry 178 to 180, stop 173.50, first target 188. Full rationale and transcript attached.

FundamentalsBuy
NewsHold
SentimentSell
TechnicalsBuy

Research, not advice

Every run is documented analysis with sources, never a recommendation to buy or sell.

Open foundations

Built on TauricResearch / TradingAgents, an open-source multi-agent framework.

Free to start

Three runs every month at no cost, and no card required to begin.

How a run works.

Every run follows the same sequence. Each step happens in order, and you can watch it happen.

  1. 01

    Pick a ticker and a date.

    AAPL on 2026-05-31, NVDA on the day of earnings, anything listed. Choose the model tier and how deep the debate should go.

  2. 02

    Watch the analyst team debate live.

    Four analysts file reports in parallel. A bull and a bear take opposing positions. A trader drafts a plan. A risk team pushes back. You see every turn as it streams.

  3. 03

    Receive a structured decision.

    BUY, HOLD, or SELL with confidence, a written rationale, and the full transcript saved to your history. About five minutes on the fast profile, ten on deep.

Inside the desk.

Twelve agents work one ticker in three stages. Every report and rebuttal is saved to the transcript.

Four analysts open the case.

Fundamentals, news, sentiment, and technicals each read the ticker independently and file a written report with its sources.

See a sample run
Fundamentals Analyst
AI

Balance sheet & valuation

News Analyst
AI

Catalysts & headlines

Sentiment Analyst
AI

Crowd positioning

Technical Analyst
AI

Price structure & momentum

Watch every turn as it streams.

No black box. Four analysts file in parallel, a bull and a bear trade rebuttals, and the desk only calls it once all twelve agents have reported.

See a sample run

Sample replay. Real runs take about five to ten minutes.

run / NVDA / 2026-06-06 / deepsample

    What you actually receive.

    Every run produces eight artifacts. They stay in your run history, exportable, with timestamps and the exact prompts used.

    Analyst reports

    • Fundamentals

      Revenue growth 8.2% YoY. Gross margin 46.1%. Free cash flow ahead of consensus by 4.3%.

    • News

      12 stories scored. Two material: supplier contract extension, antitrust filing in EU.

    • Sentiment

      Reddit and X sentiment net positive. Institutional flow neutral. Options skew bearish 30D.

    • Technicals

      Above 50D and 200D MA. RSI 62. MACD flat. Resistance at 184.20.

    Debate and plan

    • Bull vs. bear debate

      Bull: pricing power intact. Bear: channel checks softening. Three exchanges, transcript saved.

    • Trader plan

      Entry zone 178-180. Stop 173.50. First target 188. Position size 1.2% of book.

    • Risk team review

      Aggressive: lean in. Conservative: halve size. Neutral: hold to plan. Vote logged.

    Final decision

    BUY · 0.71

    Improving fundamentals outweigh near-term sentiment risk. Full rationale and the complete transcript travel with the decision.

    Sample output, not a live run.

    Runs are metered in credits and you are only charged when a run completes. Pricing is provisional during beta.

    Free

    $0forever
    Credits
    3 credits per month
    Models
    Claude Sonnet 4.6
    Start free

    Pro

    most chosen
    $29per month
    Credits
    20 credits per month
    Models
    Sonnet 4.6, GPT mini, Gemini Flash
    Start free

    Quant

    $99per month
    Credits
    70 credits per month
    Models
    Opus 4.7, GPT 5.4, full model menu
    Start free

    Put your next idea through the desk.

    Run a full multi-agent debate on any ticker in minutes. Three runs a month are free, no card required.

    Start free