Jev ecosystem projects — page 68

open-Jev LM arm: Qwen2.5-0.5B + LoRA reproducing a hosted decision model's judgment at 92.9% on hand-labelled gold - trained overnight on a 6-vCPU CPU-only host, $0/call. Paper, corpora, harnesses, receipts.

Open source implementation of Poke

Find the AI agent runs that broke because their environment did: missing keys, tools, files, permissions, network, or context. Powered by TypeSafe Jev.

Jev-powered context compaction for Pi. Keep critical instructions and tool history, prune the noise, and fall back gracefully.

Pi extension that uses Jev task classification (via OpenRouter) to route each run to explicitly configured models with fail-open policy. Public Preview.

Jev decides what to keep. Compaction never rewrites the transcript.

Jev test bench — Texas Hold'em edition: live-fire testing of TypeSafe's Jev evaluation model through a React poker game (Vercel AI Gateway). MIT.

Semantic PR gate: .jev.yml rules as TypeSafe Jev questions. Not a review bot.

Typed decisions from a System One model, where uncertainty is something you have to handle.

X（旧Twitter）のタイムラインから誹謗中傷を取り除く Chrome 拡張。判定を TypeSafe の System One モデル Jev で行う。

Static shell script analysis with Jev.

Semantic firewall for LLM agents: tool calls gated by TypeSafe Jev (System One decision model via OpenRouter) + deterministic policy. PoC with corpus, stability eval, baseline, results.

Rust CLI powered by Jev from TypeSafe.ai that ranks agent skills for the next step using live session context. Includes Claude Code hooks, structured JSON, abstention, and local feedback. Requires a TypeSafe API key.

Every tool call your agent makes, checked before it runs. A Claude Code plugin that uses TypeSafe AI's Jev to verify each pending tool call against the session plan, then allows it, asks you, or blocks it. Proof of concept

Recorded comparison of Jev, Terra, and Opus on 100 synthetic security-triage cases, five passes each, with a static inspectable dashboard

Alert triage as a UNIX filter: JSONL security alerts in, typed decisions out. Runs on TypeSafe Jev or a local model; policy routing stays in code.

A Claude Code hook that asks whether the decision you are writing needs a model at all. Includes a measured 149-row comparison of TypeSafe Jev against Claude Haiku 4.5.

MCP server exposing TypeSafe (Jev/System One) to the fleet: judge, rerank, systemone

Unofficial C++23 client for the TypeSafe AI API.

Go SDK for TypeSafe AI.

Unofficial async Rust SDK for the TypeSafe AI System One API: typed questions via #[derive(QuestionSet)], a port of typesafe-sdk-python.

typesafe-sdk-js and typesafe-sdk-python port for swift

C# Library to use the Typesafe APIs, Direct conversion from Python/JS libraries