User repository on GitHub
Jev (TypeSafe AI) PoC through Game of Thrones
- Forks
- 1
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
CommunityJev Demos & Experiments
Jev Got
Game of Thrones roleplay where a narrator model writes each scene and Jev reads it back on five questions that drive the header, music and next prompt.
- Stars
- 2
- Language
- TypeScript
- Published by
- Community
- Author
- phureewat29
- Last verified
- Oct 2026
About this project
A Game of Thrones roleplay where you play Jon Snow. Each turn a story model writes the next scene, and then TypeSafe's Jev reads that scene back and answers five questions about it: where Jon now stands, what kind of scene it was, how much danger he is in, what should play under it, and whether the prose stayed inside the fiction.
Quoted from the project's README. Read the full README.
Works forJev for Classification
Used inJev for Software & DevTools
- Languages
- TypeScript
Highlights
- Jev never writes anything: it returns a location id from a closed set, a distribution and a score, and the interface is a function of them.
- Your action is passed to the narrator as delimited data, never as instructions, and inFiction checks the result and regenerates once.
- Turn numbers are client-supplied, so a mismatch answers 409 turn_conflict and keeps two open tabs honest without a lock.
- Caps at 15 turns per session, 200 characters per action, 10 turns per minute and 60 per IP per day, plus a process-wide daily cap.
Watch out
Two keys are required, one for TypeSafe and one for OpenRouter, because a narrator model writes the prose while Jev only labels it; whatever you type is sent to both. It also needs a Redis instance for stories, so this is not something you can open and read statically. Stories expire after an untouched stretch of days by default, and turn 15 is narrated as a closing chapter. The setting belongs to George R. R. Martin and the project calls itself a non-commercial technical demonstration. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
Just the interface. A narrator chat model writes the scene, then Jev answers five questions about it: where Jon now stands, what kind of scene it was, how much danger, what should play underneath, and whether the prose stayed inside the fiction. The header, soundtrack, artwork and the next turn's prompt are all functions of those five answers.
Your action is passed to the narrator as delimited data rather than as instructions, and the inFiction judgment checks what came back, regenerating once if it falls below 0.5. Attempts to talk to the narrator are redirected in-world rather than refused, which is a deliberately different behaviour from a hard block.
It needs both a TypeSafe key and an OpenRouter key, since the prose comes from a chat model and the labels from Jev. Spend is bounded by design: 15 turns per session, 200 characters per action, 10 turns per minute and 60 per IP per day, plus a process-wide daily turn cap, and an untouched story expires on its own after the configured number of days.
The client sends the turn number it has seen and the server compares it, answering 409 turn_conflict on a mismatch, which is how two open tabs are kept honest without a lock. All model calls happen server-side so the browser never sees a key, but anything you type is sent to TypeSafe and OpenRouter.
The narrator defaults to openai/gpt-5.6-luna through OpenRouter and Jev defaults to jev-latest. Both are overridable with OPENROUTER_MODEL and TYPESAFE_MODEL, and pinning TYPESAFE_MODEL is recommended if you tune thresholds against one version.
GET /api/story/:id returns the transcript so far or the prologue for a new session without writing anything. POST /api/story/:id/turn takes an action and turn number and returns position, mood, beat, danger, text, turn, turnsRemaining, and ended. GET /api/health returns ok.
The text you type is sent to both TypeSafe and OpenRouter, and all model calls happen server-side so the browser never sees a key. The README warns not to put anything private in your action, since your action is passed to the narrator as delimited data.
More Jev resources like this
98GitHub stars
Self-hosted Docker service and web UI that packages Laya structured-decision inference behind a compatible System One endpoint.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
1Panel-dev/laya-serverPython
1Panel-dev/laya-serverPython20GitHub stars
General-purpose System One agent where Jev chooses tools and arguments while Jot executes them and loops until it can reply.
Jev Demos & Experiments#community#awesomejev#catalog-import
Community
runta-dev/jotTypeScript
runta-dev/jotTypeScript23GitHub stars
Browser drone-fleet lab where TypeSafe Jev steers each drone as fast reflexes and an optional planner offers one-use strategy advice when confidence drops, without pausing control.
Jev Demos & Experiments#community#jevbest#catalog-import
23GitHub stars
Chrome extension that sends each X post to Jev with six typed questions in one call near the viewport, then renders a behavior chip under the post plus a session cost and latency panel.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
oso95/x-scannerTypeScript
oso95/x-scannerTypeScript14GitHub stars
Manifest V3 Chrome extension in TypeScript that sends each LinkedIn post to Jev and collapses bait, self-promo, off-topic, and sponsored posts into a one-line bar with reasons and scores.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
stefw/lkcleanTypeScript
stefw/lkcleanTypeScript11GitHub stars
Standalone Flue example that routes support chat through Cloudflare Workers AI, using gpt-4o-mini for chat and Jev for typed intent and urgency routing.
Jev Demos & Experiments#community#awesomejev#catalog-import
Community
matthewp/flue-jev-demoTypeScript
matthewp/flue-jev-demoTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.
Jev Demos & Experiments#gaming#agents#control
Read article
From the Jev community
Posts from builders shipping with Jev right now.