Launches on September 17, 2026

Guardrail + model router for LLM gateways on TypeSafe's Jev (System One model), with an independent accuracy/calibration/latency evaluation. Stdlib Python.

A sandbox where a TypeSafe System One model presses the controls of a small creature. Code runs the world.

Browser Tetris with a first-class AI API: play it with a built-in planning agent, TypeSafe's Jev decision model, or any OpenAI-compatible chat model

Charts: TypeSafe Jev evaluated on Thai standardized exams vs 110 other models

Integrate decision models (Jev, OpenJev, Sage) in Rust

AI token-savings research notes + example prompts (presentation + team playbook)

Arcade games

TrustGate — indie media T&S gate via TypeSafe Jev

Unofficial CLI + Python estimator of tokens, cost and context limits for TypeSafe (System One / Jev) API requests. Not affiliated with TypeSafe.

3D chess powered by TypeSafe AI (Jev). AI vs AI by default, or play either side. Multiple difficulty levels.

Small Python package that uses typesafe.ai to evaluate code comments on certain heuristics

Diffusion-style pixel art out of a general classifier (TypeSafe Jev): 256 parallel pixel questions + refinement passes

TypeSafe.AI Go SDK

Unofficial Go SDK for the TypeSafe AI System One API — 1:1 parity with the official JS and Python SDKs. Not affiliated with TypeSafe AI.

Unofficial PHP SDK for the TypeSafe AI System One API — 1:1 parity with the official JS and Python SDKs. Not affiliated with TypeSafe AI.

Three composable judgment pipelines on TypeSafe's Jev: support-ticket triage, observability alert triage, and a deploy-risk gate.

Using Typesafe.AI to generate diff reviews.

Local document search that can refuse.

A runtime assurance layer for tool-using AI agents: it observes an execution agent independently, combines hard safeguards with Agentic Assurance, and makes the final control decision through a deterministic policy.

Extensible action gates and decision routing for AI agents, with paired cost and reliability evaluations.

Autonomous Robinhood trading agent: evidence-gated, shadow-first, with a one-click Windows app. It has to earn the right to trade, and still asks before it spends.

[Flue](https://flueframework.com) Tesrt

ASSAY-001: independent, pre-registered verification of TypeSafe Jev's calibration and type-safety claims. Split verdict, published in full.

OrthoCode AI is an evidence-first orthopedic medical-coding platform. Its runtime is constrained to a normalized, versioned codebook release and retains page-level evidence for every proposed coding line.

A production-oriented BEAM/OTP autonomy kernel for running an untrusted coding worker inside a disposable Linux execution domain while keeping durable authority, policy, effects, verification, recovery, and audit state in a trusted control plane.

Axorbis is an evidence-first AI research workspace for exploring questions, organizing sources, verifying claims, running experiments, and turning research into reliable outputs.

Routes a thread's first message to the right harness and model with TypeSafe (Jev), then asks you to confirm.

A Valheim companion bot: a BepInEx plugin that plays a real Valheim character, taking natural-language orders in game chat.

A tug-of-war over one shared actuator — you and TypeSafe's Jev push the same CartPole and the forces add. Physics verified against Gymnasium, with fair-baseline benchmarks and token accounting.

Claude Code Router

Let LLMs play Chess against each other

Watch TypeSafe's Jev classifier race across Wikipedia in real time — no LLM in the loop. Built by @dave_xt.

Claude Code hook: classify every tool call with TypeSafe's Jev (allow / ask / deny) so Claude never spends tokens judging its own tool calls

Sidecar: grade a cloudforge student writeup with TypeSafe Jev. Not part of the OSS product.

Three.js robotaxi game where TypeSafe's Jev model makes the driving decisions

A fully audited, reproducible decision‑theoretic mixture‑of‑agents framework comparing multiple selector strategies, including OpenJev, with transparent calibration, limitations, and verification artifacts.

Autonomous evolutionary trading-agent research lab for Solana. Built by Cobalt.

Distraction-free YouTube learning feed powered by TypeSafe AI's Jev System One model

Ground News style bias check for the article in your current tab.

Jev-style verbatim context compaction plugin for Hermes Agent

Jax-1-Abel: open-source typed decisions, C++ training/inference and local TypeSafe SDK gateway. Experimental distilled head, not RLCD.

An unofficial Go client for TypeSafe's System One Jev model

Agent plugin: TypeSafe Jev browser automation skills for Grok Bot / Cursor

Pre-LLM catalog gate: TypeSafe/Jev scores tools & skills; keep only high-probability items

Consumer research explorations using TypeSafe's Jev

A local lab for seeing what TypeSafe's Jev judgment model does on realistic business cases: typed answers, probabilities, policy in code, receipts.

Experiments with Typesafe's Jev

TypeSafe Jev experiments via Vercel AI Gateway: pixel B&W/gray images + long-context relevance filter bench

Sandbox & benchmark: optimize how you ask Jev (TypeSafe System One) to play falling-block puzzle games, head-to-head against LLMs

Binary LLM-judge bench — compare Jev (TypeSafe System One) against a frontier LLM judge on speed, cost, and agreement with human labels.