Bookmark Collection Rule Classifier

Krzysztof Duda tested Jev to decide whether a bookmark belongs in a collection based on its owner’s rules. Jev returns a probability: above 0.85 it files the bookmark, below 0.35 it drops it, and in between it asks the user to decide.

View on X

Krzysztof Duda@krzysztoffduda𝕏

Got access to Jev and tested it on a simple use case: does this bookmark belong in this collection, given the rules its owner wrote? Before: An LLM model answered with a sentence I had to parse into a boolean, taking 1–3s per call. Now (with Jev): It returns a probability: Above 0.85: Files it automatically. Below 0.35: Drops it. The middle: Goes back to the user as a judgment call. A sweep over 5

Sep 17, 2026X postsView on X

Previously, an LLM returned a sentence that Duda had to parse into a boolean.

Also filed under Triage & routing

Jev-Assisted Codex Model and Effort Routing
The project provides Jev-assisted model and reasoning-effort routing for local Codex Desktop and CLI.
Plain-English Gmail Filter Judge
The project uses Jev (TypeSafe) to judge Gmail filters written in plain English.
Framework-Agnostic Jev Intent and Tool Router
Huzeyfe_dev built a lightweight, framework-agnostic intent and tool router using TypeSafe’s Jev decision model to speed up agent loops.
MCP Tool Router with Jev and MCTS
Harness-router is a lightweight tool router that uses Jev for fast tool selection and Monte Carlo Tree Search for harder multi-step decisions. It uses MCP as the tool interface and falls back when routing is uncertain.