JevBystander
Zero-dependency Android app that judges the visible WeChat message's intent, emotion, urgency and reply posture with Jev.
Overview
JevBystander is an Android accessibility app for WeChat one-to-one chats. With a tap on a quick-settings tile, accessibility shortcut or notification button, it reads the messages already visible in the current chat window and sends one batched Jev request. The result is shown as exactly three toasts: intent (a 10-way Choice, such as greeting or small talk at 85%), an emotion distribution across nine options, urgency as a 0 to 3 Score, and a suggested reply posture from 11 options. A local contact table maps aliases to relation, gender and notes and is passed as state, so the same sentence is judged differently for a partner than for a colleague.
It deliberately does nothing else: no generated reply, no filling the input box, no clicking or sending, no screenshots or OCR, no reading other apps and no other uploads. The APK is 861 KB with zero third-party dependencies (hand-written JSON, HttpURLConnection and native views). The README credits jev-chat-jarvis as inspiration while stating no code was copied, and names a sister Xposed-based project by the same author. CI builds debug-signed APKs on every push. Documentation is in Chinese.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What is the other person's intent (10 options)? | — |
| Choice | What emotion does the message convey (9 options, as a distribution)? | — |
| Score | How urgent is it (0-3)? | 0 · 1 · 2 · 3 |
| Choice | What reply posture is suggested (11 options)? | — |
How it works
- User triggers via tile, shortcut or notification
- Read visible WeChat chat text through Accessibility
- One Jev request with contact relation as state
- Show three toasts; take no other action
Strengths
- Never writes or sends anything
- 861 KB with no third-party dependencies
- Relation-aware judgments via local contact table
Caveats
- WeChat one-to-one chats only
- Only sees text visible on screen
- Debug-signed builds by default
Getting started
Download JevBystander-debug.apk from the latest release (Android 8.0+), enable the accessibility service and add your TypeSafe key.At a glance
- Type
- App
- Status
- beta
- Language
- Java
- Runtime
- Android app
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Messaging: Android accessibility app that reads the visible WeChat chat window and answers one batched request of typed `Choice`, `Score` and `Boolean` questions to sort the peer's message into intent (10 options), an emotion distribution (9 options), urgency (0-3 `Score`) and a suggested reply posture (11 options), then shows exactly three toasts and takes no other action - no generated reply, no input injection, no screenshot or OCR, with a local alias-to-relation table passed as state so the same sentence is judged differently for a partner than for a colleague.
yibie/awesome-jevAndroid accessibility app that reads the visible WeChat chat screen and sends one batched Jev request (10-way intent `Choice`, 9-way emotion distribution, 0-3 urgency `Score`, 11-way reply-posture `Choice`) to show exactly three toasts - no generated reply text, no input injection, no screenshot or OCR; a local contact table supplies relation aliases as state context.
AnotiaWang/awesome-decision-modelsThis Android accessibility app reads visible WeChat one-to-one chat text and uses Jev to judge intent, emotion, urgency and reply posture, showing the result as 3 Toasts without generating or sending replies.
logicrw/awesome-jev-projects安卓无障碍版微信判读：只读屏、只弹 3 条 Toast（意图 / 情绪 / 着急 / 建议），不生成回复文案、不发送 · 零第三方依赖，APK 861 KB.
hellogumbo/awesome-jevAndroid accessibility reader for WeChat that judges the other person's message and shows three short prompts: it writes no replies and modifies nothing. Chinese.
MrJev/awesome-jev
Frequently asked questions
Does it reply for me?
No. It only shows three toasts with its judgments.
Does it take screenshots?
No, it reads only the accessibility text tree.
Why does the contact table matter?
The relation is passed as state so the same message is judged differently depending on who sent it.
Related projects
DeepSeek Harness plugin that classifies each user message's emotion, urgency and department with Jev and injects it as context.
PluginbetaTypeScriptAndroid WeChat assistant where Jev judges the chat and scores the relationship, and a chat model drafts three replies.
AppbetaKotlinWeChat plugin that classifies a long-pressed message's intent, emotion, urgency and suggested reply posture with Jev.
PluginbetaJavaLocal chat analyser that labels each WeChat message's emotion and intent with Jev and grades how your replies landed.
ProjectprototypeJavaScript- 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 Android chat co-pilot that reads the active chat via Accessibility and returns typed verdicts on intent, emotion and risk; never replies.
AppbetaKotlin