jev-chat-windows
Unofficial fork of jev-chat-windows: WeChat overlay with local OCR and drafted replies, adding optional auto-send and a knowledge base.
Overview
This is an unofficial, modified fork of jev-chat-windows, documented in Chinese, explicitly not affiliated with or endorsed by the original authors. It is a floating Windows overlay that captures your own WeChat window, reads the conversation with local offline OCR, sends recent messages to a model to draft one to three candidate replies, and fills the chosen one into the WeChat input box. The Jev judgment core and question design come from the upstream Android and Windows projects.
Additions include optional auto-send with separate switches for direct chats, group mentions and all group messages (all off by default), and a knowledge base of handwritten notes and contacts that is included in analysis. The README stresses that auto-send is irreversible because WeChat has no recall API, that the tool is for personal learning only and must not be used for mass messaging or harassment, and that it does not reverse-engineer, hook or inject into WeChat or read its database. Source is MIT, but release builds bundle a GPLv3 UI component, so the packaged zip is under GPLv3 terms.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What is the intent of the incoming message?(inferred) | — |
| Choice | Which drafted reply should be offered or sent?(inferred) | — |
How it works
- Capture your WeChat window and OCR locally
- Send recent messages for judgment and drafting
- Show one to three candidates
- Fill into WeChat; optionally auto-send if enabled
Strengths
- Local OCR without hooking WeChat
- Optional knowledge base for context
- Auto-send disabled by default
Caveats
- Unofficial fork
- Auto-send is irreversible and risks account problems
- Release package subject to GPLv3 via a UI dependency
Getting started
Download the release zip from the fork's Releases page and run the exe.At a glance
- Type
- Hook
- Status
- beta
- Language
- Python
- Runtime
- Windows desktop overlay
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Windows 微信回复助手（非官方修改版）：本地离线 OCR 读屏 + 模型起草候选回复 + 一键填入微信输入框，可选自动发送。不 hook、不注入、不读微信数据库。
hellogumbo/awesome-jev
Frequently asked questions
How is this different from jev-chat-windows?
It is an unofficial fork adding optional auto-send and a knowledge base.
Is auto-send on by default?
No, all auto-send switches are off by default.
What licence applies?
Source is MIT, but the packaged release is bound by GPLv3 through a bundled UI component.
Related projects
- jev-chat-windows★ 802
Windows WeChat reply assistant: local offline OCR reads the chat, Jev judges intent and tension and ranks three drafted replies to fill in.
ProjectproductionPython AstrBot plugin where Jev decides whether the bot should join a group conversation; the main model still writes the reply.
ProjectproductionPythoniOS custom keyboard that shows a copied message's intent, risk level and candidate replies inside any chat app, filling but never sending.
AppbetaSwiftWindows desktop chat assistant that OCRs a selected chat region, has Jev analyse the conversation, and ranks drafted replies for copying.
ProjectbetaPython, TypeScriptmacOS chat overlay that reads the current chat window and shows the incoming message's intent, risk and candidate replies, read-only.
ProjectbetaPythonWeChat reply assistant fork: DeepSeek drafts three replies, Jev pre-judges, reviews and ranks them; auto-send is cancellable.
ProjectbetaPython