All projects

The official Jev product site, introducing System One and its decision-oriented interface.

An introduction to Jev: turn state and typed questions into structured answers for your code.

The official path from the Playground to your first HTTP request.

Understand Choice, Score and Noul, and when to use each.

Request and response contracts for direct integration with the System One HTTP API.

Official architecture patterns, including confidence-gated workflows.

An official cookbook for asking multiple independent questions about the same state.

TypeSafe's introduction to System One and Jev.

The official TypeScript and JavaScript API client from TypeSafe.

The official Python client with synchronous and asynchronous interfaces.

A TypeSafeClient-compatible adapter that uses other LLM APIs as backends.

A community .NET client for System One's typed question-and-answer interface.

An official Agent Skill for designing and implementing TypeSafe decision workflows.

Jev-based model and reasoning-strategy routing for each turn in an existing Codex Router setup.

A Claude Code context tool that uses Jev to retain, truncate or remove tool records.

A community Codex port that restores original context selected by Jev around session compaction.

A community Skill covering Jev question design, state, confidence thresholds and diagnostics.

A Go CLI and MCP server that expose TypeSafe decisions to MCP-compatible agents.

A community MCP server with classification, scoring, checking and batch-question tools.

Build versioned decision functions for reuse through HTTP backends or MCP clients.

A Browser Use agent: Jev selects actions and page elements; a text model generates input when needed.

A macOS interface-control experiment that extracts screen and UI information before Jev selects actions.

An Android agent built on Mobilerun that uses Jev for action decisions.

A community Skill for Codex browser workflows: Jev chooses actions while Codex plans and verifies.

An observable browser stealth game: Jev decides guard behavior while deterministic code manages the world.

A Doom agent driven by structured spatial state, running on Chocolate Doom via WebAssembly.

A two-Jev Gomoku playground with move-by-move decisions, replays and live matches.

A Super Mario Bros. agent driven by structured emulator state.

A Snake experiment: code generates legal actions and facts; Jev selects movement directions.

A MuJoCo drone simulation connecting a small decision model to a control workflow.

A simulated Franka arm where Jev maps English goals to predefined action primitives.

A PostgreSQL extension using Jev to evaluate natural-language conditions on table data.

Community n8n nodes for System One's typed questions in automation workflows.

A community integration bringing Jev decisions to Home Assistant Assist.

Uses Search1API and Jev to interpret queries, select sources and rank results.

A .NET and React ticket-routing application using Jev for structured decisions.

Independent community React components and UI patterns for TypeSafe AI projects.

A community playground for Jev use cases, games and comparative question experiments.

TypeSafe's workflow evaluation page, presenting its methods and results.

Official notes on behavioral boundaries and failure modes in Jev 1.13.

A coding-agent safeguard that uses session context to assess tool-call and prompt-injection risks.

A community directory covering SDKs, integrations, agents, games and articles.

An independent community list of Jev use cases, projects, SDKs and learning resources.

A community ecosystem directory with source notes, categories and GitHub synchronization workflows.

A Jev project directory with a reusable GitHub review workflow.

A community showcase linking Jev projects to author sources and case descriptions.