Small decisions. Interesting possibilities.Submit contentSubmit

Jevinik

A stock research terminal that turns live market evidence into a typed 30-day outlook.

Added to Jevfast

How it uses Jev

Jev evaluates a stock ticker against dated evidence on a 30-day horizon. It returns a price-up Boolean, an overall 0–4 outlook score, an evidence-quality score, and bullish/neutral/bearish signals for each evidence category. The application exposes this as a classification result; the repository also contains an LLM-based alternative engine.

What Jev decides

Illustrative 30-day stock classification requestExample answers · not a recorded Jev response · Source ↗
Question 1 · priceUp
YOUR APP
INSTRUCTION

Based on all supplied evidence, will the ticker's share price be higher 30 days after the as-of date?

STATE

Ticker, request-date as-of date, 30-day horizon, and evidence grouped by category. Each evidence item contains a title, source, date, and snippet.

JEV · BOOLEAN
YesNo
Question 2 · outlook
YOUR APP
INSTRUCTION

Give the overall 30-day share-price outlook, weighing all evidence.

STATE

Ticker, request-date as-of date, 30-day horizon, and evidence grouped by category. Each evidence item contains a title, source, date, and snippet.

JEV · SCORE
04
Question 3 · evidenceQuality
YOUR APP
INSTRUCTION

Rate how complete, recent, and ticker-specific the evidence is.

STATE

Ticker, request-date as-of date, 30-day horizon, and evidence grouped by category. Each evidence item contains a title, source, date, and snippet.

JEV · SCORE
02

App workflow

  1. Assemble evidence

    The application groups evidence for a ticker by category, retaining each item's title, source, date, and snippet.

  2. Build the request

    It sets the as-of date to the current UTC date, fixes the horizon at 30 days, and asks for the overall Boolean, outlook and quality scores, and category signals.

  3. Classify

    The selected decision engine runs either Jev or the configured LLM and maps the structured answers into a classification result.

Why it is interesting

The request separates an overall forecast from evidence quality and per-category signals. That makes the judgment inspectable by evidence type while keeping the final outlook tied to a fixed horizon.