OPEN-SOURCE EQUITY DISCOVERY ENGINE

Find what deserves
deeper research.

RigorGate narrows a broad U.S. equity universe into a small, auditable research queue. It combines market regime, price structure, business quality, valuation, catalysts, evidence freshness, and explicit risk gates.

See how it works ↓ Try the Replay Lab

7research layers

6hard gate classes

1evidence trail

0orders created

A ranking tells you where to look. An audit trail tells you why.

From thousands of stocks to a small, inspectable research queue.

01 / DISCOVERY ENGINE

How a stock reaches
the research queue.

Each layer can narrow, pause, or reject a candidate. Missing critical evidence fails closed instead of being converted into false confidence.

01 UNIVERSE

Market coverage

U.S.-listed common stocks enter with exchange, security type, price history, and liquidity context.

START BROAD

02 CONTEXT

Regime + tradability

Market breadth, volatility, liquidity, and history gates remove structures the engine cannot underwrite.

FAIL CLOSED

03 DISCOVERY

Price structure

Trend stage, momentum, relative strength, volume, and volatility surface technically relevant setups.

RANK, DO NOT PREDICT

04 UNDERWRITE

Business + valuation

Financial durability, accounting quality, valuation context, catalysts, and sector-aware checks test the thesis.

LOOK FOR CONTRADICTIONS

05 OUTPUT

Research queue

A posture, hard-gate results, missing evidence, invalidation reference, and red-team case travel together.

HUMAN DECISION NEXT

MARKETRegime, breadth, volatility

TECHNICALStage, momentum, relative strength

BUSINESSGrowth, margins, cash conversion

ACCOUNTINGQuality, dilution, balance-sheet risk

VALUATIONContext, expectations, downside

CATALYSTWhy now, what can change

EVIDENCESource, freshness, missing data

02 / RESEARCH OUTPUT

Not a buy list.
A queue for underwriting.

The public demo shows the shape of the output using deterministic fictional securities. A live run uses configured market and filing providers.

Loading research queue…

Research candidate

Strong enough to begin full human underwriting. It is not an order.

Wait for proof

A named confirmation or missing evidence blocks advancement.

Rejected

A hard gate failed and the reason remains visible in the record.

03 / INTERACTIVE REPLAY

Step inside the
decision engine.

Use three fictional cases to experience the final decision layer. Commit before seeing the posture, then inspect what advanced or failed.

Loading replay…

—

FICTIONAL DATA

Close—

63D return—

RSI 14—

Stage—

ATR—

YOUR DECISION

Where should this case go next?

04 / EVIDENCE PASSPORT

A decision you can diff.

Every case carries posture, missing evidence, gates, invalidation, and a machine-readable observation ID.

evidence-passport.jsonREAD ONLY

{
  "status": "waiting_for_your_decision",
  "score_is_probability": false,
  "order_created": false
}

Share the evidence.
Not the hype.

Download the exact fictional replay record and inspect what the system knew, what it did not know, and why the gate held.

05 / OPEN-SOURCE CHALLENGE

Don’t just star it.
Try to break it.

Choose the smallest path that fits you. A researcher can submit a decision-time counterexample without code, a first-time contributor can ship a bounded fix, and an experienced developer can harden the evidence layer.

CODE PATH · YOUR FIRST 60 SECONDS $ make lab $ make test $ git checkout -b break-a-gate

Read the challenge rules and contribution paths ↗

OPEN SOURCE / APACHE 2.0

Build the research engine
you would trust with bad news.

Explore the repository ↗ Pick a first issue ↗

Find what deservesdeeper research.