Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

jev-helper

Browser extension that lets Jev, local Laya or an OpenAI-compatible model take over a running ra2web (WannaFire) RTS match.

Projectbeta
View on GitHub ↗

Overview

jev-helper is a Chrome and Edge Manifest V3 extension, documented in Chinese, that takes over a running match of the browser RTS ra2web (WannaFire) with one hotkey, without a Node proxy, console code or rebuilding the game. In the Jev and Laya modes, the extension presents candidate actions and the model makes a Choice among them; the README notes custom services must implement Jev's candidate Choice protocol rather than plain chat completions.

Version 0.7.5 adds features on top of the original 0.3.0: a bilingual observation panel, an in-game status overlay, three model sources (Jev cloud with a key and per-call billing, free locally deployed Laya that can be shared on a LAN, or any OpenAI-compatible service), a local decision log exportable as JSON, attack thresholds based on actual production, campaign objectives, background operation and ship self-defence, per-match records, and a commander mode where an OpenAI-compatible model sets squad tactics and build orders while the extension handles execution and reflexes. A connection test sends one candidate-choice request with no game action.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich candidate game action should be taken now?—

How it works

  1. Configure a model source and hotkey in the extension
  2. Start AI control in a match
  3. Extension presents candidate actions; Jev or Laya picks one
  4. Execute, log decisions and record match statistics

Strengths

  • No game rebuild or proxy needed
  • Free local Laya option
  • Decision logs and match records

Caveats

  • Jev cloud mode is billed per call
  • Chinese documentation
  • Game-specific

Getting started

Load the release ZIP as an unpacked extension, set the model source and hotkey (default Alt+Shift+J), then start control in a ra2web match.

At a glance

Type
Project
Status
beta
Language
TypeScript
Runtime
Chrome/Edge extension (Manifest V3)
Has code
yes
Has tests
yes

Use cases and integrations

How curated lists describe it

Frequently asked questions

What does Jev decide?

Which candidate in-game action to take.

Can it run without paying?

Yes, with a locally deployed Laya model.

Which game does it support?

ra2web, also known as WannaFire.

Related projects

  • jev-gamepilot

    Game agent claiming to combine local Laya reflexes with Jev strategy to play PC and Android games via screen capture and real inputs.

    ProjectprototypePython
  • Laya vs Jev arena

    Browser arena where local Laya and hosted Jev race in Snake and fight in a Mortal-Kombat-style game, with every move a real decision.

    ServiceprototypePython
  • laya-vs-jev

    Chrome T-Rex game where local Laya on MLX and hosted Jev choose moves side by side, with latency charts and crash replays.

    ServiceprototypePython
  • Pacman AI Race

    Browser Pac-Man race comparing hosted Jev and self-hosted Laya choosing routes from the same board state, after code removes fatal routes.

    ServiceprototypeJavaScript
  • decider

    Open family of System One-style models fine-tuned from Qwen3.5 and Gemma that return typed decision probabilities in one pass.

    ModelbetaPython
  • laya-mlx

    Independent MLX port of Laya for Apple Silicon: typed decisions in about 13 ms (7 ms multilingual) with no PyTorch or cloud API.

    ServicebetaPython