Jev 101
Start Guides
Search
中文 EN
Ops heat 1 · Quiet★ 0heat incomplete (no X)

Called It!

A MIT Claude Code / Codex skill that grades annual predictions at strict, standard and lenient bars, using hosted TypeSafe Jev for calibrated probability grades over committed evidence rows.

Repo: rickmanelius/called-it ↗

What it is

Grade predictions with receipts, not vibes.

What it does: Builds three bars per claim, gathers public evidence, and asks hosted Jev whether each bar is met so hit rates stay replayable.

Why it belongs in the showcase: End-to-end TypeSafe Jev grader with committed inputs and offline replay.

Caveats: Optional Jev key; bars are public opinion.

Evidence

Public MIT repo https://github.com/rickmanelius/called-it (default branch master), created 2026-10-27, 0 stars as of 2026-10-29. README stages Discover→Extract→Operationalize→Gather→Grade with TypeSafe Jev (TYPESAFE_API_KEY optional; host model answers same typed questions without a key). Companion article https://www.rickmanelius.com/p/called-it (HTTP 200) and gallery https://rickmanelius.github.io/called-it/. Offline replay with no API key.

Caveats

  • Without TYPESAFE_API_KEY the host model grades the same typed questions (less calibrated)
  • Bars and evidence tiers are author-defined; argue via PR, not by editing scores in place
  • X hydration and paywalled sources need optional tokens or a browser

Unofficial case page. Stars / list position observed for heat; do not treat heat as a quality SLA. Review the upstream README before you depend on it.