Safe GitHub Repositories (2026) · Open Source Tools

Semantic tool routing and typed System One decisions for the Pi coding agent using TypeSafe Jev

An unofficial local visualizer for Jev (TypeSafe): a live view of every call your code makes. Not affiliated with TypeSafe AI.

Thinking coreutils

Claude Code / Codex / pi plugin that hands agent steps needing no text output to Jev (TypeSafe's judgment model) — measured p50 ~230 ms and ~$0.02 per 1,000 judgments, with typed escalation back to the LLM

ACP and MCP adapter that bridges TypeSafe Jev with any LLM — computer use and typed decisions alongside Codex, Claude, Grok, and OpenCode.

Go SDK and CLI for TypeSafe Jev: typed decisions (yes/no, choice, score) from a model

TypeSafe SDK in Swift

Police detector for Android: spots police body cams, patrol cars, Flock ALPR cameras, AirTag trackers, Remote ID drones, smart glasses and fake cell towers from Bluetooth LE and WiFi. Free, open source, no root.

The decision layer for your Rails app. A Rails-native wrapper around TypeSafe's Jev System One API: typed, calibrated decisions in your control flow.

Semantic decisions as ordinary Ruby #feels?, #decide, #score, Rails validations and pattern matching powered by Jev

A calibrated context sieve for Claude Code: every tool result is judged by a System One model before it enters context.

Type-safe, framework-agnostic, composable, and tree-shakeable TypeScript libraries for the web.

Open framework for confidential AI

Cone Programming Language

A staged code-review workflow and local dashboard built with TypeSafe Jev.

Private photo vault for Android