jev-charactor
TypeScript framework for role-playing NPCs: Jev picks actions and communicative intents, the host composes dialogue.
Overview
Jev-Character is a modular TypeScript framework for role-playing agents and autonomous game NPCs powered by Jev. Developers describe character principles in natural language and supply changing observations and memories; Jev selects eligible actions and communicative intents, while the host executes world effects and composes responses from authored plans and permitted facts. This supports interactions between characters and their environment, between characters, and with the player, without free-form text generation in the Jev dialogue path. It combines agent memory, character knowledge and personal goals in a reusable SDK.
Fog Harbor is the playable case study: a bilingual 2D AI town with five residents and a player trying to make a living, demonstrating a small multi-agent simulation. The README is explicit that the project is a software alpha and not evidence of large-scale social emergence or of superiority over rule-based or LLM-based agents, and frames it as exploring a path from individual characters toward agent societies.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which eligible action should this character take now? | — |
| Choice | What communicative intent should this character express? | — |
How it works
- Author character principles in natural language
- Supply observations and memories each tick
- Jev selects an action and intent
- Host executes effects and composes dialogue from authored plans
Strengths
- No free-form text generation in the dialogue path
- Reusable character SDK
- Playable demo town
Caveats
- Alpha software
- No claims of emergent social behaviour
At a glance
- Type
- Project
- Status
- prototype
- Language
- TypeScript
- Runtime
- library and browser game
- Has code
- yes
Use cases and integrations
How curated lists describe it
Jev-Character: a TypeScript core for role-playing agents, autonomous game NPCs and AI towns. Agent memory, natural-language behavior authoring and programmatic dialogue, with a playable 2D multi-agent simulation.
logicrw/awesome-jev-projects
Frequently asked questions
What does Jev decide for characters?
Which eligible action to take and which communicative intent to express.
Who writes the dialogue?
The host composes it from authored plans and permitted facts.
What is Fog Harbor?
A bilingual 2D AI town demo with five residents.
Related projects
AI town simulation forked from a16z ai-town where Jev picks each agent's next legal move and an LLM handles dialogue and memory.
AppprototypeTypeScriptReal-time bot battles where you write a fighter's strategy in plain English and a System One model pilots it several times a second.
ProjectbetaTypeScriptBot that plays Red Alert 2 web skirmishes, with code handling mechanics and a decision model making stance, threat and timing calls.
ServicebetaPythonMultiplayer 3D science-fantasy version of Houston's River Oaks District with Jev-powered NPC companion, chauffeur and smart drive.
ProjectbetaTypeScriptBrowser 3D planetary rover sandbox where a TypeSafe controller chooses the next action from complete candidates, versus a baseline.
ProjectbetaTypeScriptDoom runs as a spacecraft payload over real flight software; Jev pilots from the ground every half second while Claude reviews.
ProjectprototypePython