User repository on GitHub
切る牌コーチ — 1人用麻雀コーチ (Vercel AI Gateway Jev + Turso)
- Forks
- 1
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Demos & Experiments
Kiru Hai Coach
A solo mahjong coach that takes a 14-tile hand and has Jev recommend which tile to discard, with efficiency, safety and wait scores plus a Japanese explanation, beginner-to-advanced lesson tracks, and a practice mode against three CPUs. Progress and consultations live in Turso.
- Stars
- 0
- Language
- TypeScript
- Reading time
- 2 min
- Published by
- Community
- Author
- smilior
- Last verified
- Oct 2026
Used inJev for Gaming
- Languages
- TypeScript
- Platforms
- Next.jsVercel AI Gateway
- Capabilities
- Open sourceRequires API key
Highlights
- From a 14-tile hand Jev recommends one tile to discard and explains it in Japanese with efficiency, safety and wait scores.
- Lesson tracks run beginner to advanced; V1.5 adds practice against three CPU players in a landscape-only table layout.
- Built on Next.js App Router and TypeScript with AI SDK's experimental_evaluate and the typesafe-ai/jev model through Vercel AI Gateway.
- Progress and consultation history live in Turso, keyed by an anonymous player_id stored in cookie or localStorage.
- Tile images are CC0 from FluffyStuff/riichi-mahjong-tiles; application code is not licensed for reuse.
Quickstart
bash
npm install
cp .env.example .env.local
npm run devWatch out
Grounded in the complete Japanese-language README and GitHub metadata reviewed on 2026-10-15; in-game decisions for CPU players are heuristic while only the human seat consults Jev, and the application code carries no open licence. 0 stars; the tile art is CC0. Live at kiru-hai-coach.vercel.app.
Questions about this entry
Short answers drawn from the linked project or page.
A solo mahjong coach: from a 14-tile hand, Jev recommends which tile to discard and explains the choice in Japanese using efficiency, safety and wait scores. It is live at kiru-hai-coach.vercel.app.
The /api/coach route posts a hand of 14 tile IDs plus difficulty and optional river tiles to Jev through the Vercel AI Gateway, using AI SDK's experimental_evaluate with the typesafe-ai/jev model, and returns scores and commentary.
npm install, copy .env.example to .env.local, and set AI_GATEWAY_API_KEY, TURSO_DATABASE_URL and TURSO_AUTH_TOKEN. A setup script creates the Turso database, or a migration script updates an existing one; then npm run dev.
Beginner, intermediate and advanced lesson tracks plus a V1.5 practice mode against three CPU players, with simplified rules without yaku, scoring or calls, where only the human seat's moves are explained by Jev.
Turso is the source of truth for progress, lessons and consultation history. An anonymous player_id kept in a cookie or localStorage restores your session when you return, and scores are stored as efficiency, safety and wait.
You need a Vercel AI Gateway key and a Turso database, practice mode is landscape-only, and the README publishes no accuracy claims for the coaching. CPU opponents use heuristics rather than Jev to keep calls down.
smilior built it; it had 0 stars and 1 fork at review time, and the README says application code belongs to the repository owner rather than an open licence. The mahjong tile art is CC0 from FluffyStuff/riichi-mahjong-tiles.
More Jev resources like this
A Next.js console whose whole design system is generated at runtime by Jev from a one-sentence brief: type 'an 18th century printed book' and every visual decision is chosen, not written, from a catalog a person authored.
Jev Demos & Experiments#community#catalog-import#madewithjev
Community
bilune/jev-designTypeScript
bilune/jev-designTypeScriptTurkish-language natural-language search over an outerwear catalog: Jev reads each request in one call and scores candidate matches, while code handles price and size regex, filtering, ranking and the results UI — with a side panel showing what Jev decided.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
selmakcby/jev-kiyafet-bulTypeScript
selmakcby/jev-kiyafet-bulTypeScriptA Next.js and MUI showcase of Jev classifying Enron-Spam emails: one call asks a Noul spam question, a Choice category question and a Score urgency question at once. Includes a dataset view, a no-key example tab and a stand-alone Python script.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
harsha89/jev-demoTypeScript
harsha89/jev-demoTypeScriptJevSpeak explores whether a model that cannot generate text can still talk: about 13 parallel Jev decisions per turn become a typed Semantic IR that a deterministic compiler renders in English or Chinese. Live at jevspeak.org.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
mm-sheng/jevspeakTypeScript
mm-sheng/jevspeakTypeScriptA three-level AI interrogation game powered by TypeSafe Jev System One: write a statement, face the warden's verdict built from live structured analysis of plea, logic and paradox, and try to earn your release. Runs on Bun with SQLite locally and Neon/Postgres in production.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
bugkiwi/turing-jailTypeScript
bugkiwi/turing-jailTypeScriptA browser chess coach where Stockfish (WASM) owns the truth about the position and Jev makes the coaching judgment after every move, with templates and a small writing model supplying the words. Status: Wave 0 scaffold and Jev feasibility spike.
Jev Demos & Experiments#community#catalog-import#madewithjev
Community
joellewis/game-coachTypeScript
joellewis/game-coachTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.