jev-chat
Android chat copilot: Jev judges intent, needs and risk in visible chats, then ranks drafted replies to fill, never send.
This summary is based on how curated lists describe the project; its README was not available to review.
Overview
jev-chat is an Android chat copilot described in several lists. An accessibility service reads the visible conversation in apps such as QQ, X, Feishu/Lark or WhatsApp (Feishu text via on-device OCR), and one batched Jev request asks Choice questions about the other person's intent, needs and the next action, a Score for danger, and Noul checks on subtext, substance and tension. A separate chat model drafts three replies, one more Jev Choice ranks them, and the app fills the picked reply into the input box.
Sending always stays manual. The hellogumbo summary describes it as non-intrusive: it only reads the screen, without hooking or modifying the apps. No README was fetched for this repository, so setup, permissions and data handling details are not confirmed here.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What is the other person's intent and what do they need?(inferred) | — |
| Score | How dangerous or risky is this conversation?(inferred) | — |
| Choice | Which of three drafted replies is best?(inferred) | — |
How it works
- Accessibility service reads the visible chat
- Jev answers intent, needs, next-action, danger and subtext questions
- A chat model drafts three replies
- Jev ranks them; the best is filled into the input box
- User decides whether to send
Strengths
- Never sends on its own
- Works across several messaging apps without modifying them
Caveats
- Reads private conversations via accessibility service
- No README available to verify details
At a glance
- Type
- Hook
- Runtime
- Android app (accessibility service)
Use cases and integrations
How curated lists describe it
Messaging: an Android accessibility service reads the visible QQ, X, Feishu/Lark or WhatsApp conversation, one Jev request asks `Choice` questions on intent, needs and next action, a `Score` for danger and `Noul` checks on subtext, substance and tension, then one more `Choice` ranks three replies drafted by a separate chat model, and the app fills the pick into the input box while sending stays manual.
yibie/awesome-jevUnofficial Android chat copilot: an accessibility service reads the visible QQ, X, or Feishu/Lark chat (Feishu text via on-device OCR), one batched Jev request asks intent, needs, and next-action `Choice` questions, a danger `Score`, and three `Noul` checks, a separate chat model drafts three replies that one more `Choice` ranks, and code fills the picked reply into the input box without sending
AnotiaWang/awesome-decision-modelsThe chat decision assistant: before you reply, Jev reads the chat, judges…
wh000wh000/awesome-jev-liveAndroid chat copilot where Jev answers intent, needs, and next-action Choice questions, a danger-level Score, and three Noul checks about the visible QQ, X, or Feishu/Lark conversation, then a Choice ranks three replies drafted by a separate chat model; code fills the picked reply into the input box and never sends.
cobanov/awesome-jev装在手机上的对话副驾：在微信 / QQ / X / 飞书里读懂对方、给出候选回复、一键填入输入框，发不发由你。非侵入，只读屏幕，不 hook 不改包。
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
The other person's intent, needs, risk level and which drafted reply to suggest.
Does it send messages?
No, it only fills the input box; sending is manual.
Which apps does it support?
QQ, X, Feishu/Lark and WhatsApp per the lists.
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 Android chat co-pilot that reads the active chat via Accessibility and returns typed verdicts on intent, emotion and risk; never replies.
AppbetaKotlinSelf-hosted helper for tricky chats and everyday decisions: Jev makes typed checks, an LLM handles language, the user decides.
ProjectprototypeTypeScriptDeepSeek Harness plugin that classifies each user message's emotion, urgency and department with Jev and injects it as context.
PluginbetaTypeScriptZero-dependency Android app that judges the visible WeChat message's intent, emotion, urgency and reply posture with Jev.
AppbetaJavaAndroid WeChat assistant where Jev judges the chat and scores the relationship, and a chat model drafts three replies.
AppbetaKotlin