Jevthoven
A music studio where Jev selects musical patterns and code turns them into editable MIDI.
What Jev decides
Integration evidence ↗Chooses instruments, harmony, and bar patterns from bounded options.
Loading repository and market data…
Loading repository and market data…
Discover what builders are making with Jev, from game-playing agents to music and everyday automation.
Jev selects musical patterns. The app renders those choices into notes you can edit, play, and export. Explore how a decision model becomes a creative instrument.
Explore the projectA music studio where Jev selects musical patterns and code turns them into editable MIDI.
Chooses instruments, harmony, and bar patterns from bounded options.
Typed decisions for shell scripts, batch jobs, and coding-agent workflows.
Answers yes/no, choice, and score questions supplied by the command line.
A Snake experiment with a decision model choosing each move.
Selects among legal moves using board facts calculated by the game engine.
A game-playing agent that works from structured emulator observations.
Chooses gameplay actions from the emulator state.
Home Assistant integration with support for hosted Jev and compatible decision APIs.
Evaluates typed automation questions, including boolean, choice, and numeric scores.
A set of MCP tools that brings Jev judgments into coding agents.
Checks claims against evidence, ranks candidates, and reviews proposed changes.
An experimental coding assistant built around bounded Jev workflows.
Routes requests to workflows and answers fixed-choice questions about code evidence.
OpenTelemetry log triage that identifies useful signals before deeper analysis.
Scores diagnostic value and urgency to route logs while preserving the archive.
Worked examples for building applications with Jev's typed decisions.
Turns application state into choices, boolean judgments, and numeric scores.
A prototype case-management board for cross-border payment operations.
Classifies, prioritizes, and routes incoming cases to rules and human review.
An active-learning pipeline for allocating training-data labeling work.
Accepts high-confidence labels and queues uncertain examples for a teacher model or people.
Opted in by repository maintainers using the gitpad-jev topic. These integrations have not been reviewed and do not receive a Uses Jev badge.
No community submissions indexed yet. Add your project →
Independent community directory. “Uses Jev” means the linked project documentation describes an integration; it is not an audit or an endorsement by TypeSafe. A community market is separate from the project and does not imply owner approval.