xrtm v0.9

AI for event forecasting

Run probabilistic forecasts from the command line. Uses your existing API key. Works with any OpenAI-compatible endpoint.

terminal — xrtm demo

How it works

Install

pip install xrtm. No GPU, no Docker. Set your API key in .env and go.

Forecast

xrtm start runs a linear pipeline: load questions → forecast → score → backtest → report.

Inspect

xrtm runs show --latest prints Brier scores, reasoning traces, and all artifacts.

xrtm start runs a linear pipeline: questions → forecast → score → report. All in a single command.

One provider, any endpoint. Works with any service that speaks the OpenAI Chat API.

Brier score, ECE, log score. Full reasoning traces with causal DAGs. 8+ artifacts per run.

pip install xrtm-forecast for composable agents, topologies, and providers in Python.

Built-in 21-question corpus. Live markets via Polymarket and Metaculus.

Apache 2.0. 5 packages. Python 3.11+. No GPU. No Docker.