A collection of markdown skill files that ground AI agents in Android development best practices, installable via the Android CLI.
jev-chat/jev-chat-jarvis
A chat decision assistant that reads the conversation, judges intent and risk, and drafts replies you send yourself on Android, Windows, macOS and iOS.
- Stars
- 7,923
- Forks
- 1,287
- Open issues
- 30
- Last commit
- 4 days ago
What is it?
jev-chat is a chat assistant that analyzes the other person's intent, the risk in the moment and the best response before drafting any reply. It then produces ranked reply options and fills the one you pick into the message box. It never sends messages on its own.
The project ships as several apps. On Android, an accessibility-based overlay works with WhatsApp, QQ, Feishu and X. On Windows and macOS, the app reads a chat window using screenshots with on-device OCR or system text APIs. The iOS version is a custom keyboard that works on copied messages and is distributed as source to build with Xcode.
According to the project, Android chat text is sent only to a model provider the user configures with their own key, and screenshots used for OCR stay on the phone. The code is MIT licensed.
Key features
- 01Judges the other person's intent and risk before drafting replies.
- 02Ranks drafted replies and fills the chosen one into the input box.
- 03Works as an overlay on WhatsApp, QQ, Feishu and X on Android.
- 04Reads chat windows on Windows and macOS via screenshots and on-device OCR.
- 05Requires you to press send; it never sends messages automatically.
- 06Uses your own model provider key, with no project-run server.
Use cases
- Deciding how to reply to a tense or ambiguous message on WhatsApp or QQ.
- Getting suggested replies for work chats in Feishu on Android.
- Reviewing a copied message on iPhone through a custom keyboard before answering.
Quick start
- 1
Download the Android APK for your chats from the Releases page
- 2
Install the app, then follow its guide to turn on the accessibility service and overlay permission
- 3
Configure your own model provider API key in the app
Things to know
- Android use requires granting accessibility service and overlay permissions.
- The iOS keyboard has no App Store build and must be built from source with Xcode.
- The macOS app supports Apple Silicon only.
- Using it inside apps like QQ, Feishu, X or WhatsApp may breach their terms of service and risk account restrictions.
Topics
FAQ
- What is jev-chat-jarvis?
- A chat decision assistant that reads the conversation, judges intent and risk, and drafts replies you send yourself on Android, Windows, macOS and iOS. jev-chat is a chat assistant that analyzes the other person's intent, the risk in the moment and the best response before drafting any reply. It then produces ranked reply options and fills the one you pick into the message box. It never sends messages on its own.
- Is jev-chat-jarvis free to use? What license does it have?
- jev-chat-jarvis is open source under the MIT license. Check the license terms before using it in commercial products.
- Is jev-chat-jarvis still maintained?
- Actively maintained. The last commit to jev-chat-jarvis was on Oct 16, 2026.
- What can I use jev-chat-jarvis for?
- Deciding how to reply to a tense or ambiguous message on WhatsApp or QQ. Getting suggested replies for work chats in Feishu on Android. Reviewing a copied message on iPhone through a custom keyboard before answering.
- How do I install jev-chat-jarvis?
- Download the Android APK for your chats from the Releases page Install the app, then follow its guide to turn on the accessibility service and overlay permission Configure your own model provider API key in the app
- Where can I get jev-chat-jarvis?
- The source code and installation instructions are on GitHub: https://github.com/jev-chat/jev-chat-jarvis
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An Android image editor app with dozens of tools, from crop and draw to 500+ filters, OCR, background removal, and PDF tools.
Offline, privacy-first PDF toolkit for Android that merges, splits, compresses, converts, signs and annotates PDFs entirely on-device.
A lightweight personal AI assistant written in Go that runs on low-cost hardware and connects to many LLM providers and chat apps.
An Android and iOS app from Google AI Edge for trying open-source LLMs and on-device GenAI use cases offline on your phone.
An agent framework that connects digital humans (2.5D, 3D, mobile, PC, web) and LLMs (OpenAI-compatible, DeepSeek) to business systems.