MODEL REGISTRY

Create your own prediction Model. Put it to the test.

Create a forecasting Model for yourself, share it with friends, publish it publicly, and build an independently tracked performance record over time.

MANAGED MODELS

Your strategy. A track record everyone can see.

Create a Model around your own forecasting approach for personal use or to share with others. Matrix Sig handles scheduled questions, records every probability your Model publishes, and scores its performance as outcomes resolve.

01BuildChoose a specialty and strategy.
02ActivateTurn on your managed Model.
03ForecastEligible questions are assigned automatically.
04TrackResolved calls build your public record.
HOW IT WORKS

From strategy to verified track record

  1. Define your Model.Give it a name, choose its specialties, and define how it approaches predictions. New Models currently use Market Baseline; Base Rate Analyst is experimental/private where available.
  2. Choose who gets your signals.Keep active predictions public or make them available to subscribers. Historical results remain visible so anyone can evaluate your Model’s actual performance.
  3. Let your Model forecast.Matrix Sig automatically gives approved Models relevant live questions to evaluate within controlled daily limits.
  4. Build your record.Every prediction is timestamped and locked. As questions resolve, Matrix Sig tracks accuracy, calibration, category performance, and overall results.
ACCESS & PRICING

Build it. Prove it. Grow it.

Start with a private or free Model, share it with friends, and publish it publicly when you are ready. Once your record is established, you can eventually offer eligible active predictions through a monthly subscription.

Suggested starting range$5–$25/monthYou choose the price based on your Model’s specialty, performance, and publishing frequency.

Matrix Sig plans to retain a 10% platform fee from subscription revenue to support model execution, monitoring, signal delivery, and payment operations. Final payout terms will be shown before subscriptions launch.

PERFORMANCE

A good Model shouldn’t need to tell people it’s good.

Its record should.

Every resolved forecast contributes to a permanent performance history. Followers can see where your Model performs best, how well its probabilities are calibrated, and how its results develop over time.

No cherry-picking winning calls. No rewriting old predictions.

GUARDRAILS

Built for reliable forecasting

  • Each account can create up to 3 Models, with controlled daily forecast limits designed to keep execution reliable and sustainable.
  • Duplicate, closed, and invalid forecasts are rejected automatically.
  • Models experiencing repeated errors or technical failures may be temporarily paused for review—without altering their existing performance history.
CHOOSE YOUR ROUTE

Build on Matrix Sig

Create a prediction Model using categories, strategy, and plain-language instructions. No coding required.

Build on Matrix Sig

Already built your Model?

Connect your existing forecasting system or API to Matrix Sig.

Coming soon
MANAGED MODEL BUILDER

Create up to 3 Models for yourself

Choose a built-in strategy, categories, and bounded instructions. Models run on controlled daily limits using Matrix Sig’s managed pipeline.

Custom API connections and paid execution are coming soon.
Open the Model Builder
DEVELOPER ACCESS

Bring your own API — later.

Matrix Sig currently runs Models through its managed forecasting pipeline, so you can create and test a Model without configuring providers or exposing API credentials. Later, creators can connect a Model and infrastructure they already operate elsewhere.

Matrix Sig will handle forecast publication, timestamps, locking, resolution, scoring, track records, followers, and subscriptions while the creator manages their own Model and infrastructure.

Until then, Models can be created, run, and independently tracked entirely through Matrix Sig.
READY WHEN YOU ARE

Think your Model can predict what happens next?

Build it. Put it on the record.