Jev Gallery
中文EN

Updates

A record of actual collection changes. Follow updates with RSS.

Subscribe via RSS →

2026-10-15

Four application additions and a Jev 1.14 guide

Added jevcrawl, jevgate, Jev Agent Control and Mastra Competitor Monitor for focused crawling, command gates, agent handoff and competitor-page changes. A bilingual release guide distinguishes preview from latest and covers experimental images, Box, long context and SDK migration. Official documents and pinned source were reviewed; upstream models were not run.

2026-10-12

Four Jev additions: page ads, skill loading, email and project triage

Added Jev Ad Block, JevShelf, Jev Email Triage and pm-jev for webpage judgments, lazy skill loading, two-stage email decisions and project-item triage. Pinned source and licenses were reviewed without running upstream apps or models; prototype status, optional Jev backends and outbound-data boundaries are explicit.

2026-10-09

Three new Jev cases, plus an updated Home Assistant integration

Added a Jev-driven Pac-Man game, an evaluation of Convex coding knowledge, and a Claude Code skill-routing experiment with a negative field result; the Home Assistant entry was refreshed against current source. The catalog now has 151 items (including eight demos without repositories). Pinned source and licenses were reviewed; projects and models were not independently run, and author-reported results retain their stated scope.

2026-10-07

Four new cases: danmaku filtering, shell search, character motion and evaluation

Added Jev Danmaku Filter, Compass, Open Annie and jev-eval. The catalog now contains 148 cases, including eight demos. Pinned source, Jev integrations and licenses were reviewed without independent model runtime tests.

2026-10-05

Four new cases: email, jobs, products and a simulated city

Added Postmark, AI PM Job Radar, Catalog Match Lab and Jev Plays SimCity. The catalog now contains 144 cases, including eight demos. Pinned source, Jev integrations and MIT licenses were reviewed without independent model runtime tests.

2026-10-01

Three new cases and ten previously prepared additions

Reviewed and added Chess with Jev, 1 Million Emojis and Jev Chat Gate. Ten previously prepared cases are also published, bringing the catalog to 140 cases including eight demos. New additions have source and license reviews, without independent model runtime tests.

2026-09-30

Five new Jev application and research cases

Added code navigation, DuckDB semantic queries, Griptape workflows, novel character profiles and an HN problem-analysis prototype. The catalog now contains 137 cases, including eight demos. All additions have pinned-source and license reviews; none was independently runtime-tested.

2026-09-27

Five new practical Jev applications

Added a classification API, Paperless filing, IMAP sorting, PR triage and an information radar. The collection now contains 132 cases, including eight demos. All five additions have pinned source and license reviews; none was independently runtime-tested.

2026-09-26

Five new Jev applications and demos

Added voice spellcasting, browser-history retrieval, test selection, a village simulation and a command-palette demo. The gallery now has 127 cases, including eight demos. Three additions have pinned-source reviews; two have original-post and sampled-video evidence. None was independently runtime-tested.

2026-09-25

Demo collection launched with PokéJev

A dedicated bilingual demo collection brings together six application demos without linked public source, including the new PokéJev natural-language Pokémon filtering video. The gallery now has 122 cases, retaining use categories and explicit evidence boundaries.

2026-09-25

Four new Jev applications

Added natural-language filters, mobile-device navigation, Python MCP evidence checks and Codex tool recommendations. The collection now has 121 cases. Pinned source and licenses were reviewed without paid model calls or runtime replication.

2026-09-24

Six new Jev applications

Added customer-agent decision checks, qualitative coding, Facebook feed filtering, GitHub Actions, Pi effort selection, and App Store keyword research. The collection now has 117 cases. Pinned source was reviewed without live model calls or independent runtime testing.

2026-09-23

Five new Jev workflow and creative applications

Added tmux pane assessment, n8n review routing, audience-directed video stories, command-risk assessment and agent evaluations. The collection now has 111 cases. Pinned source was reviewed without model calls or independent runtime testing.

2026-09-22

New: ChatJEVs, a Jev chat assistant

Traced a user-shared X post to the original author and repository. Reviewed the Android assistant’s assessment, separate reply generation, ranking and manual sending flow. The collection now has 106 cases, with remote processing and the lack of independent runtime testing disclosed.

2026-09-21

Eight new Jev cases, including a reviewed submission

Eight projects created September 20–21 cover demo trading, YouTube comments, browser traces, HN discussions, WordPress links, semantic file search, logistics triage and 2048. The collection now has 105 cases, all source-reviewed without independent execution or paid inference.

2026-09-20

Four new workflow guides

New guides cover Playwright semantic testing, Feishu message triage, Python record linkage and writing quality checks. The library now offers 18 topics and 36 bilingual articles connected to 97 cases, with synthetic fixtures distinguished from source evidence.

2026-09-20

12 recently published Jev applications

This batch adds 12 applications from repositories created September 17–20: lesson notes, Feishu triage, film selection, read-only disk inspection, semantic testing, writing checks, operations, Go, sponsor skipping, privacy labels, record linkage and voice control. The collection now has 97 cases. Pinned documentation and call sites were reviewed; none was independently run.

2026-09-19

Eight new Jev application guides

The guide library now covers 14 topics, adding bilingual RAG, browser automation, moderation, n8n, citation review, transcript analysis, NPC and model-selection guides. Source-based explanations, synthetic examples and editorial recommendations are labeled; no live-model tests were added.

2026-09-19

24 more Jev applications

Video scoring, moderation, paper evidence extraction, speech prompting, command suggestions, notification filtering and games expand the collection from 61 to 85. Pinned documentation and Jev call sites were reviewed for every new project; none was independently run.

2026-09-19

New case: Jev-cu / Jev Use

Sac’s Codex Computer Use project lets Jev select the next step from UI text candidates. The author’s post and pinned source were reviewed; speed and token usage were not independently tested. The collection now includes 61 projects and demos.

View case →

2026-09-19

Easier discovery and sharing

All 60 cases now have dedicated Chinese and English pages, with search, category and evidence filters, editors’ picks and 5 editorial notes. Submission drafts survive language changes in the same tab. Independently tested case count remains 0.

2026-09-19

Six practical Jev guides

New guides cover the model, API, prompts, MCP, Claude Code context compaction and text classification, with explicit example and evidence boundaries.

Explore guides →