What is jev-chat-windows-laya?

A Windows fork of Liyucheng1997/332_lab-jev-chat, a WeChat desktop chat assistant. The fork keeps the upstream Android code untouched and changes only the Windows desktop edition: a DPI-aware capture fix plus an optional local judge backend.

What was the DPI bug and why did it matter?

On a 175%-scaled display, Win32 reported a 1042×848 logical rect while ImageGrab captured 1824×1484 pixels, so the app read the wrong screen area. Worse than broken: unrelated screen text could be sent to TypeSafe or DeepSeek.

How do I install and start it?

Clone the repo, run windows/install.ps1 to build the venv and dependencies, then windows/start.ps1. First use: open WeChat desktop, click the select-chat-region button, frame only the message bubbles, then click analyze current conversation.

Can I run it without a TypeSafe API key?

Yes, via JEV_BACKEND=laya, a local open-source judging engine where the text never leaves your machine. But the author's own five-scenario test found the base checkpoint unusable for Chinese chat judgment, so TypeSafe remains the default.

What are the measured problems with the local laya engine?

On five hand-written Chinese scenarios: multilingual mode with the project's 7 questions hit 2/5 intents with a 0.84 danger span, and one clearly angry accusation was classified as light chat with 0.94 confidence. The README recommends fine-tuning laya first.

What does the app refuse to do?

It offers copy-only reply suggestions with no auto-send path, and it refuses to analyze when it detects transfer, red-packet, receipt or payment wording. It reads only chats on your own device that you are entitled to view.

How are the fork's changes documented?

FORK-NOTES.md lists every change with evidence and measurements, including the one diagnosed-but-unfixed issue: an over-broad UIA criterion that can capture WeChat window titles. A clean fix/dpi-awareness branch covers 4 files (+166/−8) for an upstream PR.

Jev directory

CommunityJev Tools & CLIs

Jev Chat Windows Laya

A Windows fork of the jev-chat WeChat assistant that fixes DPI-unaware screen capture on scaled displays and adds an optional key-free local laya judge — with the author's own measurements showing that local engine is not yet usable for Chinese chat.

See highlights Open github.com

Stars: 4
Language: Python
License: MIT
Reading time: 2 min
Published by: Community
Author: zjemyoung
Last verified: Oct 2026

zjemyoung/jev-chat-windows-laya

User repository on GitHub

View on GitHub

Windows 版微信聊天副驾（上游 fork）：本地 laya 判断引擎免密钥运行 + 修复高缩放屏抓取错位 Windows fork of jev-chat: key-free local laya judge + DPI-aware screen capture fix

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity catalog-import logicrw windows wechat

Works forJev for Classification Jev for Scoring

Used inJev for Software & DevTools

Languages: Python
Capabilities: Open sourceRuns locallyRequires API key

Highlights

At 175% scaling the old code cropped a 1042×848 logical rect while ImageGrab captured 1824×1484 pixels, so the wrong region was read.
WeChat 4.x does not expose chat text to UI Automation, making OCR the only capture path — so the scaling bug necessarily triggered.
After the fix the selection matches, window sizing follows the scale (875×1330, tk scaling 2.332) and all 14 upstream unit tests pass.
Optional JEV_BACKEND=laya runs a local key-free judge; on five Chinese scenarios it hit 2/5 intents, one angry message read 'light chat' at 0.94.
Copy-only by design: no auto-send path, and analysis refuses when transfer, red-packet or payment wording is detected.

Quickstart

bash

git clone https://github.com/ZJemYoung/jev-chat-windows-laya.git
cd jev-chat-windows-laya
powershell -ExecutionPolicy Bypass -File .\windows\install.ps1
powershell -ExecutionPolicy Bypass -File .\windows\start.ps1

Watch out

Copy reviewed against the truncated repo README on 2026-10-15; the capture ends inside the expanded upstream README, so upstream features beyond the fork notes were not reviewed. English copy here translates the Chinese README, and the laya accuracy figures are the fork author's own five-scenario measurements.

Questions about this entry

Short answers drawn from the linked project or page.

From the Jev guides

Original write-ups on what this entry is for.

Jev at home: voice pipelines, Assist routing, and sentinels

Oct 16, 2026

How Home Assistant conversation agents, macOS voice pipelines, and an advisory sentinel put a fast typed decision in the loop — one call per utterance, thresholds in code, and a daily token budget.

Jev Tools & CLIs#home-assistant#voice#automation

Read article

Search visibility, audited by decisions

Oct 16, 2026

Jev-based SEO tools split the crawl from the call: deterministic crawlers count and parse, while typed judgments rank fixes, place internal links, score search results, and track AI-answer mentions — with costs and caveats attached.

Jev Tools & CLIs#seo#search-visibility#audits

Read article

Jev on the Gateways: Vercel, Workers AI, and OpenRouter

Oct 12, 2026

The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.

Jev Tools & CLIs#gateway#vercel#cloudflare

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai

Jev Chat Windows Laya

Highlights

Quickstart

Questions about this entry

What is jev-chat-windows-laya?

What was the DPI bug and why did it matter?

How do I install and start it?

Can I run it without a TypeSafe API key?

What are the measured problems with the local laya engine?

What does the app refuse to do?

How are the fork's changes documented?

More Jev resources like this

From the Jev guides

From the Jev AI model community