Skip to main content

Examples and proof

This page is intentionally product-and-proof-first. Everything below is either a shipped CLI workflow or a source example that exists today. Future work belongs on the roadmap.

Page role

Released proof plus source examples. Use this page to find verifiable examples that exist today. For package/API ownership, continue to Developer / integrator workflow. For unreleased-but-real features, use Next release track.

Ownership and proof

This page helps you find verifiable examples without making xrtm.org the owner of them. The newcomer CLI flow ships from xrtm-org/xrtm, package example code ships from the repo that contains it, and schema or policy changes that affect those examples start in xrtm-org/governance.

Official proof-point workflows​

The story is simple: XRTM is AI for event forecasting. These workflows and examples are the shipped proof behind that claim on the released 0.3.1 surface:

  1. Provider-free first success
  2. Benchmark smoke workflow
  3. Monitoring, history, and report workflow
  4. Local-LLM advanced workflow

1. Provider-free first success​

If you only do one thing, do this first:

xrtm demo --provider mock --limit 1 --runs-dir runs
xrtm runs list --runs-dir runs
xrtm runs show <run-id> --runs-dir runs
xrtm artifacts inspect runs/<run-id>
xrtm report html runs/<run-id>
xrtm web --runs-dir runs

That flow is the core proof that XRTM works today as a repeatable event-forecasting system. The commands are surfaced here for newcomers, but the behavior, flags, and artifact contracts are owned by xrtm-org/xrtm.

2. Benchmark smoke workflow​

xrtm perf run --scenario provider-free-smoke --iterations 3 --limit 1 --runs-dir runs-perf --output performance.json

This is the honest default baseline. Repeated mock-provider runs should stay stable, which makes later changes legible.

3. Monitoring, history, and report workflow​

xrtm profile create my-local --provider mock --limit 2 --runs-dir runs
xrtm run profile my-local
xrtm monitor start --provider mock --limit 2 --runs-dir runs
xrtm runs compare <run-id-a> <run-id-b> --runs-dir runs
xrtm runs export <run-id> --runs-dir runs --output export.json

Use this as the released compare gate:

  • unchanged mock-provider comparisons mean you still have the control run
  • lower Brier/ECE with similar warnings/errors and acceptable runtime is the real improvement signal
  • export is the follow-up step after compare, not the proof by itself

4. Local-LLM advanced workflow​

export XRTM_LOCAL_LLM_BASE_URL=http://localhost:8080/v1
xrtm local-llm status
xrtm demo --provider local-llm --limit 1 --max-tokens 768 --runs-dir runs-local

What the product ships today​

  • provider-free demo mode via xrtm demo --provider mock --limit 1 --runs-dir runs
  • provider-free benchmark smoke via xrtm perf run
  • canonical run artifacts under runs/<run-id>/
  • scored outputs including eval.json, run_summary.json, and report.html
  • local WebUI and TUI backed by the same run artifacts
  • run history, compare, search, and JSON export commands
  • profile-based repeatable workflows via xrtm profile create
  • local monitoring lifecycle commands
  • optional local-LLM health and demo commands

Examples by owning repo​

These examples are real shipped source assets, but they are examples, not the default first-run product path. When example code changes, update the repo that owns the file and refresh this page only when the navigation or framing should change.

xrtm-forecast (xrtm-org/forecast)​

Useful example entry points include:

  • examples/kit/minimal_agent/run_minimal_agent.py
  • examples/providers/provider_free_analyst/run_provider_free_analyst.py (deterministic no-key path via the top-level xrtm package)
  • examples/kit/local_analyst/run_local_analyst.py (optional local-LLM path)
  • examples/kit/pipelines/forecasting_analyst/run_forecasting_analyst.py
  • examples/kit/topologies/debate_demo/run_debate_demo.py
  • examples/kit/features/discovery/run_discovery.py
  • examples/kit/features/streaming_demo/run_streaming_demo.py
  • examples/kit/topologies/consensus_demo/run_consensus_demo.py

See the repo: xrtm-org/forecast.

xrtm-train (xrtm-org/train)​

Useful example entry points live in the external xrtm-org/train repo:

How to read examples honestly​

  • Product workflow first: start with Getting started.
  • Workflow docs second: use the researcher and operator paths for the supported control → candidate → compare journeys.
  • Examples third: reach for package scripts when you want to learn internals, extend the system, or study a narrower capability.
  • Roadmap separate: if something is only planned or experimental, it should not be treated as a hero claim.