CalibratedDecisions.

Repo · Support & inbox

jev-triage (OpenClaw)

OpenClaw plugin that puts TypeSafe Jev in front of the agent's main model: each incoming message is classified thanks/spam/question/incident plus…

Open github.com ↗

From the repository

Jev, TypeSafe AI's decision model, triages every message before your OpenClaw agent's main model runs.

How builders describe it

OpenClaw plugin that puts TypeSafe Jev in front of the agent's main model: each incoming message is classified thanks/spam/question/incident plus urgent yes/no, so thanks and spam end without waking the main model and other messages arrive with Jev's triage prepended; every decision is logged.

The decision Jev makes

Pick the destination queue, set a priority, or decide whether an item needs a human.

Where it fits

Support and operations teams route work all day: which queue, how urgent, who handles it. Jev turns each routing step into a typed choice with a confidence score, so uncertain items go to a person. All 81 support & inbox projects →

Related projects