JEVCASES / COMMUNITY INDEX
See what people
build with Jev.
Discover community projects, experiments, and demos built with Jev. Browse the cases, see what Jev decides, and find each project’s source or available demo.
- INDEXED CASES
- 44 live + featured + community
- LIVE DEMOS
- 02 open now
- SOURCE LINKS
- 42 community repositories
- CATEGORIES
- 32 ways to explore
Explore the case wall
jev-ultrafast
A browser-agent experiment where a small text model handles typing while Jev picks the operation and the observed element.
json-render
A Vercel Labs framework for generative UI, explored here with a Jev composition step that selects components, properties, and layout.
QuantDinger
An open-source trading platform that can place typed Jev entry and risk checks in front of eligible live orders.
fast-jev-compaction
Context compaction for Claude Code that judges tool calls and results, then keeps the surviving text verbatim instead of rewriting it.
jev-trader
A Monad / Kuru market-making experiment where buy and sell decisions are selected from order-book state.
agent-desktop
Desktop automation that reads OS accessibility trees so structured controls can inform decisions about buttons, menus, and inputs.
jev-review
Produces structured code-review judgments over diffs or source files, prioritizes risk, and shows the results in a local dashboard.
jev-router
A loopback proxy that lets Jev select a model tier for each fresh Claude Code or Codex user turn.
DocJev
Classifies documents or splits mixed PDF packets by category and page boundaries using Jev, with local text extraction and optional OCR.
Mobile Jev
An Android agent with a React studio and CLI: Jev chooses actions from observed controls, and Mobilerun executes them.
typesafe-mario
Turns emulator telemetry and RAM into structured state so Jev can select legal controller inputs without screenshots.
jev-mcp
A judgment toolbox for agents: verification, screening, semantic matching and ranking, classification, and extraction.
Unclutter
A browser extension that classifies page clutter with Jev and reuses reversible hiding rules across matching page templates.
typesafe-mcp
Exposes structured Choice, Score, and Noul judgments over the Model Context Protocol, connecting to compatible assistants.
jev-codex-router
Per-call model and reasoning routing for Codex: Jev picks the capability tier and thinking depth for every model call, tool continuations included.
killmyidea
Scores a startup idea across bounded dimensions and computes a KILL, FIX, or SHIP verdict.
jev-drone
A MuJoCo drone simulation where Jev makes higher-level tactical judgments while deterministic flight control and safety limits keep authority.
pi-warden
A Pi extension that checks proposed edits against project rules and asks the agent to supply evidence for completion claims.
sys1grep
A semantic grep that asks whether each line satisfies a proposition, then combines the answers with AND, OR, and NOT.
neo4jev
Navigates a Neo4j graph by choosing outgoing relationships and judging whether the goal has been reached.
SkillRanker
A Rust CLI that compares eligible skills with the current agent context and can recommend none when fit is weak.
Prism
Jev classifies market conditions such as toxic flow, pressure, and mean reversion, while strategy code and risk gates control execution.
Jevmeter
Combines transcript timing, Jev sentence judgments, and an editable video plan to render annotated clips.
Winnow
Judges large Read, Bash, and Grep results for task relevance, keeping useful or uncertain blocks and caching the hidden material for recall.
Canny
Checks completion claims against a ledger of edits and commands, where deterministic facts enforce the gates and Jev adds optional semantic judgment.
Blink
Natural-language codebase navigation where walkers select the relevant directories and files while traversing the tree.
jev-curate
Judges quality, relevance, and risk in JSONL and Parquet training records before keeping or filtering them.
SemDecide
Typed semantic decisions inside shell pipelines: predicates, classification, scoring, and filtering for CI and data workflows.
OneVOneJev
A browser 1v1 FPS where structured decision ticks choose movement, look direction, aiming, firing, and jumping.
Jev plays StarCraft
A StarCraft shareware harness that offers Jev bounded economy and combat commands, then executes ordinary game inputs.
Jev Cookbook
Runnable Jev recipes for support triage, indexing, tagging, deduplication, PII detection, extraction, search re-ranking, and browser tasks.
Jev Logs
Scores log value, priority, and actionability before an optional deeper-analysis branch while preserving the existing archive.
Jevflake
Brings Jev judgments into Snowflake SQL and dbt, with stored answers, change detection, and a review queue.
jev-mobile
A durable Android sub-agent that stores tasks in SQLite, lets Jev choose a valid next action, and verifies the result on the device.
jev-turbo
Turns mapped page controls into named candidates, then lets Jev choose the next browser action and judge completion.
Barrunto
A Chrome extension that combines typed content judgments with site-specific rules and stays quiet when evidence is weak.
jev-skill-router
A Claude Code hook that asks Jev which installed skill fits a prompt and can log or inject the suggestion in a shadow-first experiment.
psearch
A terminal and MCP search tool that uses Jev to score fetched evidence and select links for a bounded search queue.
jev-oncall
Incident triage where Jev judges typed alert questions and deterministic code routes paging, human review, or suppression.
Sift
A Chrome extension that scores organic search snippets, reorders results, and explains why a result was promoted or folded.
jev-table-import-mapper
Maps uploaded CSV columns to a destination table with deterministic matching first, then typed Jev decisions for ambiguous matches.
Jev Gmail Classifier
A Google Apps Script project that turns probability-based questions about email threads into configured labels and moves.
GitHub star snapshots: 2026-09-25 – 2026-09-27. Open a case for its check date, project context, and source.
Live demos
A clearly labeled placement for AI teams.
$20 / 30 days