Content
# 🛑
> **AI conversation "premature ejaculation" terminator, allowing communication to continue to the end**
Are you still frustrated that the AI assistant always ends the conversation early? It clearly has a lot more to talk about, but it says "Is there anything else I can help you with?" **Cunzhi** specializes in this problem!
When the AI wants to "finish things off hastily", Cunzhi will pop up a dialog box in time, allowing you to continue in-depth communication until the problem is truly solved.
## 🌟 Core Features
- 🛑 **Intelligent Interception**: Automatically pop up continue options when AI wants to end
- 🧠 **Memory Management**: Store development specifications and preferences by project
- 🎨 **Elegant Interaction**: Markdown support, multiple input methods
- ⚡ **Instant Use**: 3-second installation, cross-platform support
## 📸 See the effect
### 🛑 Intelligent Interception Pop-up Window

*When the AI wants to end the conversation, the Cunzhi intelligent pop-up window appears in time, providing predefined options for quick selection, allowing communication to continue in depth*
### ⚙️ Settings Management Interface

*Elegant settings interface, supporting memory management, function switches, theme switching, and intelligent prompt generation*
## 🚀 Get Started
### Method 1: Quick Installation (Recommended)
**macOS Users:**
```bash
# 3-second installation, no need to wait for compilation
brew tap imhuso/cunzhi && brew install cunzhi
```
> ⚠️ **Troubleshooting**: If you encounter a 404 download error, please follow these steps:
>
> ```bash
> # Clear old tap cache
> brew untap imhuso/cunzhi
> # Reinstall
> brew tap imhuso/cunzhi && brew install cunzhi
> ```
>
> **Common Error Example:**
> ```
> curl: (56) The requested URL returned error: 404
> Error: cunzhi: Failed to download resource "cunzhi"
> Download failed: https://github.com/imhuso/cunzhi/releases/download/v0.2.0/cunzhi-cli-v0.2.4-macos-aarch64.tar.gz
> ```
>
> This is usually caused by a version number mismatch due to a cache problem with the previous version. Clearing the tap and reinstalling will solve the problem.
**Windows/Linux Users:**
It is recommended to manually download using method 2, which is simple and fast!
### Method 2: Manual Download
1. Visit [Releases Page](https://github.com/imhuso/cunzhi/releases)
2. Download the version suitable for your system:
- 🐧 **Linux**: `cunzhi-cli-v*-linux-x86_64.tar.gz`
- 🍎 **macOS (Intel)**: `cunzhi-cli-v*-macos-x86_64.tar.gz`
- 🍎 **macOS (Apple Silicon)**: `cunzhi-cli-v*-macos-aarch64.tar.gz`
- 🪟 **Windows**: `cunzhi-cli-v*-windows-x86_64.zip`
3. After decompression, add `寸止` and `等一下` to the system PATH
## ⚡ Quick Start
### Step 1: Configure MCP Client
Add the following to your MCP client (such as Claude Desktop) configuration file:
```json
{
"mcpServers": {
"寸止": {
"command": "寸止"
}
}
}
```
### Step 2: Open the Settings Interface
```bash
# Run the settings command
等一下
```
### Step 3: Configure Prompts
In the "Reference Prompts" tab of the settings interface:
1. View automatically generated prompts
2. Click the copy button
3. Add the prompt to your AI assistant
### Step 4: Start Using
Now your AI assistant has intelligent interception, memory management, and pop-up interaction functions!
> 💡 **Tips**: You can refer to the generated prompts to make personalized modifications and create an exclusive AI interaction experience.
## 🔧 Tool Description
寸止 provides multiple MCP tools to enhance the capabilities of AI assistants:
- **Code Search Tool**: ACE-based semantic code search to help AI understand project code structure
- 📖 [Detailed Usage Instructions](./ACEMCP.md)
### 🙏 Acknowledgements
Thanks to the following open-source projects and their contributors:
- **[acemcp](https://github.com/qy527145/acemcp)** - Developed by [@qy527145](https://github.com/qy527145), providing powerful code library semantic search capabilities. This project rewrites the core logic using Rust and integrates it into the Cunzhi MCP tool ecosystem while retaining the original functions.
## 🤝 Contribute
寸止 is an open-source project, and we welcome all forms of contribution!
### 🛠️ Local Development
```bash
git clone https://github.com/imhuso/cunzhi.git
cd cunzhi
pnpm install
pnpm tauri:dev
```
## 📄 Open Source License
MIT License - Free to use, welcome to contribute!
</div>
Connection Info
You Might Also Like
markitdown
MarkItDown-MCP is a lightweight server for converting URIs to Markdown.
servers
Model Context Protocol Servers
everything-claude-code
Complete Claude Code configuration collection - agents, skills, hooks,...
Time
A Model Context Protocol server for time and timezone conversions.
Filesystem
Node.js MCP Server for filesystem operations with dynamic access control.
Sequential Thinking
A structured MCP server for dynamic problem-solving and reflective thinking.