11
Unofficial TypeSafe Jev showcase — System One decisions, not chat.
Research and benchmarks — page 4
Benchmarks, papers and independent reproductions. Read these when you need to understand behaviour rather than adopt a library.
537 active projects in this collection. All collections →
11
Coding-agent memory where every fact is a verbatim quote graded by TypeSafe Jev's calibrated confidence. Local-first, SQLite receipts, zero dependencies.
11
Open-source, local Jev alternative: a System One decision layer with provably calibrated confidence (ECE 0.170→0.071). Typed decisions, runs offline, no API key, no waitlist.
11
Agent skills for designing, training, evaluating and improving application-specific decision systems. Primitive/model selection, data assembly, export/reload and bounded hill climbing. TypeSafe Jev is the default hosted exemplar; independent of TypeSafe.
11
Backtest Jev (TypeSafe) as a BUY/SELL/HOLD trader on NQ L10 order-book data
Hermes Agent plans, Jev (TypeSafe) picks bounded actions, Mineflayer executes: Minecraft with no screenshots or keypresses from a model. Includes the reproduction of rmalde/minecraft-agent's Ender Dragon run.
10
A local, offline System One server compatible with TypeSafe's Jev API. It answers typed yes/no, category and score questions with small open models.
10
Stop guessing confidence thresholds: calibrate, threshold, and drift-check typed decision models (TypeSafe Jev) against an LLM teacher.
10
Reproducible calibration and selective-risk benchmarks for Jev/TypeSafe decisions in DSPy workflows
Evidence-first long-term experiment memory skill for AI coding agents, with optional Jev decision-model layers
9
⚡ Foq — the FREE, local, open-source alternative to Jev. Typed System 1 decisions in ~25 ms — no waitlist, no cloud, no per-token cost. foq.fr
SLO-aware LLM inference router with Jev decisions, live queue metrics, counterfactual evaluation, and reproducible latency/cost benchmarks
Does a TypeSafe Jev rerank beat embedding search? Graded relevance eval (9,831 pairs, 164 zh/en queries) over the Agent Skills Hub catalog, with the judge-circularity bias measured.
8
Benchmark TypeSafe JEV against LLMs, fine-tuned BERT, Laya and zero-shot NLI on text classification: accuracy, calibration, latency, throughput, cost
Find, design, and evaluate TypeSafe Jev decision loops.
8
Opensource Jev
Local Jev-compatible evaluation server: POST /v1/systemone with typed noul/choice/score, open weights, no waitlist
8
Nushell module for the TypeSafe System One API: typed decisions with calibrated probabilities
Model performance first. Token savings second. A Pi extension with freshness-aware read dedupe, Jev log filtering, and searchable verbatim recall. Keeps existing message history intact.
TypeSafe Jev baseline for DimABSA (SemEval-2026 Task 3) subtask 1: zero-shot and 3-shot valence-arousal regression
7
Luce: a recipe for calibrated decision models — a sentence about your task in, a small model that answers typed questions with honest probabilities out (init → synth → train → eval → serve)
Independent calibration test of TypeSafe's Jev on a task it cannot have seen: 900 rule-generated support tickets (choice / score / boolean) plus 3 public benchmarks via Vercel AI Gateway. Raw responses, ECE with noise floor, temperature refit, per-type sign of miscalibration. Reproducible for ~$0.06.
NanoJev Snake Arena: 1v4 human-vs-AI battleship + 100-agent swarm simulator. Local demo of Jev System-One model (open-source mini replica).
Claude Code plugin that scores review findings, debug hypotheses and design options with TypeSafe's Jev — calibrated probabilities instead of one more opinion.