Caja Inversora AI platform interface showing live analysis dashboards

Features

Every feature, built around one question: can we prove it?

Caja Inversora AI is designed so that each capability is traceable — from raw data ingestion to the recommendation you read. Below is a full walkthrough of what the platform does and why it works this way.

All figures referenced below are illustrative examples of platform mechanics, not live performance claims.

Core capabilities

What Caja Inversora AI actually does

Four pillars make up the platform. Each one is documented publicly so students and independent researchers can inspect the logic instead of trusting a black box.

Structured data intake

Market data, on-chain signals, and public sentiment feeds are pulled on a fixed schedule and normalized before anything is analyzed. Inconsistent or delayed sources are flagged rather than silently used.

Rule-based signal generation

Signals are produced from documented rule sets rather than opaque discretionary calls. Each output is tagged with the rule version that generated it, so the same input always produces a traceable, reviewable result.

Public outcome logging

Every signal is logged with a timestamp before any outcome is known, then revisited and marked once the outcome resolves. This sequence — log first, resolve later — is what keeps the record honest.

Configurable risk posture

Because every learner has a different risk tolerance, the platform separates "what the data shows" from "how aggressively to act on it," letting you set your own posture without changing the underlying analysis.

Caja Inversora AI research desk reviewing logged signal data
The same logging pipeline is used internally and shown externally — nothing is repackaged for presentation.

How it runs

From data point to dashboard entry

Each feature exists to close a specific gap left by typical retail research tools. Here is the sequence a single piece of information travels through.

  1. 01

    Ingest and timestamp

    Raw data is pulled from multiple sources and stamped with an intake time before any processing begins, preserving a clean audit trail.

  2. 02

    Apply the rule set

    A documented, versioned rule set converts raw inputs into a signal. The version number is recorded alongside the output.

  3. 03

    Publish to the log

    The signal is written to the public transparency log immediately, prior to any outcome being known, closing the door on after-the-fact editing.

  4. 04

    Resolve and mark status

    Once enough time has passed, the entry is reviewed and marked as verified or held for further review, never quietly deleted.

  5. 05

    Surface to your dashboard

    Depending on your configured risk posture, the resolved signal and its history are surfaced with context, not just a bare recommendation.

Feature terms, explained plainly

Rule version
An identifier attached to every signal indicating exactly which logic produced it, so changes over time remain traceable.
Resolution window
The fixed period after a signal is logged during which its outcome is tracked before being marked verified or under review.
Risk posture
A user-configurable setting that adjusts how aggressively suggestions are framed, without altering the underlying data analysis.
Transparency log
The public, append-only record of signals and their eventual outcomes, viewable independent of any individual account.

Adjustable by design

Two modes, one shared data set

Not every learner wants the same tone of suggestion. Caja Inversora AI keeps the underlying analysis identical across modes and only changes how conservatively that analysis is framed and surfaced.

Switching modes never changes historical log entries — past signals remain exactly as they were recorded, regardless of which mode you're currently viewing them in.

Preview: framing mode

Conservative framing favors caution and wider margins in how signals are presented; standard framing presents the same underlying signal with less hedging language. Both draw from the same logged data.

Why this matters

Most tools bury their assumptions inside a single fixed output. Separating "data" from "framing" lets you see exactly where interpretation begins.

It also means you can compare how a conservative read and a standard read diverge on the same event — a useful exercise for understanding your own risk tolerance.

This feature is a study aid, not financial advice, and every mode carries the same underlying uncertainty.

Access & support

What's included when you request access

Every feature above is bundled together — there is no separate tier that withholds the transparency log or the configurable risk posture.

Data feeds
Multi-source
Log visibility
Public
Risk modes
2 built-in
Rule updates
Versioned

Have a question about how a specific feature works before you request access? Reach out through the contact page and we'll walk you through it.