AI-powered investment research

Invest with conviction,
not consensus.

Opinionated AI agents that encode history's greatest investment minds, delivering clear theses backed by rigorous, transparent analysis.

Explore research Meet the agents
3
Investor-inspired agents
Weekly
Fresh research every Friday
8
Holdings in the current portfolio

Why opinionated beats neutral.

Most financial tools are designed to be neutral. Our agents hold clear investment philosophies, forcing transparency in reasoning and accountability in outcomes.

Opinionated by design

Each agent embodies a specific investment philosophy, delivering conviction instead of run-of-the-mill consensus.

Margin of safety

Every valuation includes a rigorous margin of safety, because the best returns come from buying below intrinsic value.

Risk-first thinking

Downside scenarios are evaluated before upside potential, protecting capital the way legendary investors do.

Data-dense insights

Clear visualizations distill complex data, from balance sheets to ownership structures.

Meet the agents.

Each agent replicates the decision framework of a legendary investor, analyzing every company through its own lens.

Value & Moat

Warren Buffett

Durable competitive advantages, consistent earnings, and management quality.

Learn more

Deep Value & Safety

Benjamin Graham

A margin of safety, conservative balance sheets, and a stable earnings history.

Learn more

Quality & Mental Models

Charlie Munger

Wonderful businesses at fair prices, judged through a latticework of mental models.

Learn more

Research

Reports that take a position.

Fully detailed weekly reports with historical metrics, interactive charts and, most importantly, the clear, conviction-driven verdicts of our agents.

Historical metrics

Fundamentals, financial ratios, proprietary scoring and actionable key insights.

Interactive charts

Ownership structures, scoring radars and interactive trend charts.

Agent verdicts

Direct theses from every agent, with a signal, a confidence level and full reasoning.

Explore research

A portfolio built on evidence.

Market equilibrium is the starting point. Confidence-weighted agent views decide where to tilt.

Current allocation

8 holdings

out of 14 candidates screened

Rebalanced

Oct 31, 2026

GGBR4 12.5%
ITUB4 12.5%
LREN3 12.5%
NATU3 12.5%
RDOR3 12.5%
RENT3 12.5%
SHUL4 12.5%
SUZB3 12.5%

Screened out

AXIA3 CSAN3 ENEV3 FLRY3 LWSA3 RAIL3

How it's built.

A Bayesian pipeline that keeps every weight traceable back to the analysis behind it.

  1. 1
    Agent views
    Each asset is judged by investor-inspired agents.
  2. 2
    Bayesian update
    Equilibrium returns are updated by confidence-weighted views.
  3. 3
    Risk-aware weights
    The most robust configuration is selected every week.
Read how the engine works
Explore the portfolio

Writings.

Essays and research on our philosophy, the science behind our engine, and the road from thesis to market.

Methodology

Inside the engine: from agent opinions to a portfolio

How specialized agents evaluate every asset, how a Bayesian framework turns their opinions into confidence-weighted views, and how dynamic allocation keeps principal at the center of every decision.

September 24, 2026 · 6 min read

Research

Grounding Investor Views: Neural Predicates in the Black-Litterman Model

Our new paper shows how neural predicates can generate Black-Litterman views and their uncertainty directly from data, making every portfolio weight traceable and the whole pipeline learnable end to end.

July 24, 2026 · 4 min read

Vision

Conviction, not consensus: why financial agents should have opinions

Investors have never had more information, and never felt more lost. The answer isn't another data feed: it's agents with a point of view.

March 24, 2026 · 5 min read
Read all writings