claude-glm-alt-installer 2.0.0 → 2.0.2

package/README.md

@@ -1,5 +1,14 @@
 # Claude-GLM Wrapper
 
+> **📢 Community Fork Notice**
+> 
+> This is an actively maintained community fork of the original [claude-glm-wrapper](https://github.com/JoeInnsp23/claude-glm-wrapper). 
+> The upstream repository is currently unmaintained. This fork includes bug fixes, new features, and ongoing support.
+> 
+> Install via: `npx claude-glm-alt-installer`
+
+---
+
 Use [Z.AI's GLM models](https://z.ai) with [Claude Code](https://www.anthropic.com/claude-code) — **without losing your existing Claude setup!**
 
 Switch freely between multiple AI providers: GLM, OpenAI, Gemini, OpenRouter, and Anthropic Claude.
@@ -23,6 +32,7 @@ npx claude-glm-alt-installer
 ```
 
 Then activate (platform-specific):
+
 ```bash
 # macOS / Linux:
 source ~/.zshrc  # or ~/.bashrc
@@ -34,6 +44,7 @@ source ~/.zshrc  # or ~/.bashrc
 ### Start Using GLM Models
 
 **All Platforms:**
+
 ```bash
 ccg              # Claude Code with GLM-4.7 (latest)
 ccg46            # Claude Code with GLM-4.6
@@ -95,6 +106,7 @@ npx claude-glm-alt-installer
 ```
 
 The installer will:
+
 - Auto-detect your operating system
 - Check if Claude Code is installed
 - Ask for your Z.AI API key
@@ -117,12 +129,14 @@ source ~/.zshrc  # or ~/.bashrc
 <summary>macOS / Linux</summary>
 
 **One-Line Install:**
+
 ```bash
 bash <(curl -fsSL https://raw.githubusercontent.com/MohMaya/claude-glm-wrapper/main/install.sh)
 source ~/.zshrc  # or ~/.bashrc
 ```
 
 **Clone and Install:**
+
 ```bash
 git clone https://github.com/MohMaya/claude-glm-wrapper.git
 cd claude-glm-wrapper
@@ -136,12 +150,14 @@ source ~/.zshrc
 <summary>Windows (PowerShell)</summary>
 
 **One-Line Install:**
+
 ```powershell
 iwr -useb https://raw.githubusercontent.com/MohMaya/claude-glm-wrapper/main/install.ps1 | iex
 . $PROFILE
 ```
 
 **Clone and Install:**
+
 ```powershell
 git clone https://github.com/MohMaya/claude-glm-wrapper.git
 cd claude-glm-wrapper
@@ -150,6 +166,7 @@ cd claude-glm-wrapper
 ```
 
 **Note:** If you get an execution policy error, run:
+
 ```powershell
 Set-ExecutionPolicy -Scope CurrentUser RemoteSigned
 ```
@@ -176,6 +193,7 @@ The installer creates these commands and aliases:
 **🆕 New: ccx Multi-Provider Proxy**
 
 The `ccx` command starts a local proxy that lets you switch between multiple AI providers in a single session:
+
 - **OpenAI**: GPT-4o, GPT-4o-mini, and more
 - **OpenRouter**: Access to hundreds of models
 - **Google Gemini**: Gemini 1.5 Pro and Flash
@@ -187,6 +205,7 @@ Switch models mid-session using `/model <provider>:<model-name>`. Perfect for co
 ### How It Works
 
 Each command starts a **separate Claude Code session** with different configurations:
+
 - `ccg`, `ccg45`, and `ccf` use Z.AI's API with your Z.AI key
 - `cc` uses Anthropic's API with your Anthropic key (default Claude setup)
 - Your configurations **never conflict** — they're stored in separate directories
@@ -194,36 +213,42 @@ Each command starts a **separate Claude Code session** with different configurat
 ### Basic Examples
 
 **Start a coding session with the latest GLM:**
+
 ```bash
 ccg
 # Opens Claude Code using GLM-4.7
 ```
 
 **Use GLM-4.6:**
+
 ```bash
 ccg46
 # Opens Claude Code using GLM-4.6
 ```
 
 **Use GLM-4.5:**
+
 ```bash
 ccg45
 # Opens Claude Code using GLM-4.5
 ```
 
 **Need faster responses? Use the fast model:**
+
 ```bash
 ccf
 # Opens Claude Code using GLM-4.5-Air
 ```
 
 **Use regular Claude:**
+
 ```bash
 cc
 # Opens Claude Code with Anthropic models (your default setup)
 ```
 
 **Pass arguments like normal:**
+
 ```bash
 ccg --help
 ccg "refactor this function"
@@ -233,6 +258,7 @@ ccf "quick question about Python"
 ## Common Workflows
 
 ### Workflow 1: Testing with GLM, Production with Claude
+
 ```bash
 # Develop and test with cost-effective GLM-4.7
 ccg
@@ -245,6 +271,7 @@ cc
 ```
 
 ### Workflow 2: Quick Questions with Fast Model
+
 ```bash
 # Quick syntax questions
 ccf "how do I use async/await in Python?"
@@ -255,6 +282,7 @@ ccg
 ```
 
 ### Workflow 3: Multiple Projects
+
 ```bash
 # Project 1: Use GLM to save costs
 cd ~/project1
@@ -337,6 +365,7 @@ Use Claude Code's built-in `/model` command with provider prefixes:
 ### ccx Workflows
 
 **Workflow 1: Compare Model Responses**
+
 ```bash
 ccx
 # Ask a question
@@ -347,6 +376,7 @@ ccx
 ```
 
 **Workflow 2: Cost Optimization**
+
 ```bash
 ccx
 # Start with a fast, cheap model for exploration
@@ -357,6 +387,7 @@ ccx
 ```
 
 **Workflow 3: Leverage Model Strengths**
+
 ```bash
 ccx
 # Use GPT-4 for coding
@@ -395,6 +426,7 @@ ccx
 Each wrapper uses its own configuration directory to prevent conflicts:
 
 **macOS / Linux:**
+
 | Command | Config Directory | Purpose |
 |---------|-----------------|---------|
 | `claude-glm` | `~/.claude-glm/` | GLM-4.7 settings and history |
@@ -404,6 +436,7 @@ Each wrapper uses its own configuration directory to prevent conflicts:
 | `claude` | `~/.claude/` (default) | Your original Claude setup |
 
 **Windows:**
+
 | Command | Config Directory | Purpose |
 |---------|-----------------|---------|
 | `claude-glm` | `%USERPROFILE%\.claude-glm\` | GLM-4.7 settings and history |
@@ -413,6 +446,7 @@ Each wrapper uses its own configuration directory to prevent conflicts:
 | `claude` | `%USERPROFILE%\.claude\` (default) | Your original Claude setup |
 
 **This means:**
+
 - ✅ Your original Claude settings are **never touched**
 - ✅ Chat histories stay separate for each model
 - ✅ API keys are isolated — no mixing!
@@ -420,12 +454,14 @@ Each wrapper uses its own configuration directory to prevent conflicts:
 ### Wrapper Scripts Location
 
 **macOS / Linux:** `~/.local/bin/`
+
 - `claude-glm` (GLM-4.7)
 - `claude-glm-4.6` (GLM-4.6)
 - `claude-glm-4.5` (GLM-4.5)
 - `claude-glm-fast` (GLM-4.5-Air)
 
 **Windows:** `%USERPROFILE%\.local\bin\`
+
 - `claude-glm.ps1` (GLM-4.7)
 - `claude-glm-4.6.ps1` (GLM-4.6)
 - `claude-glm-4.5.ps1` (GLM-4.5)
@@ -438,12 +474,14 @@ These are just tiny wrapper scripts (bash or PowerShell) that set the right envi
 ### macOS / Linux
 
 **Option 1: Use the Installer**
+
 ```bash
 cd claude-glm-wrapper && bash install.sh
 # Choose option "1) Update API key only"
 ```
 
 **Option 2: Edit Manually**
+
 ```bash
 nano ~/.local/bin/claude-glm
 nano ~/.local/bin/claude-glm-4.6
@@ -455,6 +493,7 @@ nano ~/.local/bin/claude-glm-fast
 ### Windows (PowerShell)
 
 **Option 1: Use the Installer**
+
 ```powershell
 cd claude-glm-wrapper
 .\install.ps1
@@ -462,6 +501,7 @@ cd claude-glm-wrapper
 ```
 
 **Option 2: Edit Manually**
+
 ```powershell
 notepad "$env:USERPROFILE\.local\bin\claude-glm.ps1"
 notepad "$env:USERPROFILE\.local\bin\claude-glm-4.6.ps1"
@@ -490,6 +530,7 @@ Claude Code reads these variables and uses them instead of the defaults. Simple!
 **Problem**: Claude Code isn't installed or not in your PATH.
 
 **Solutions**:
+
 1. Install Claude Code from [anthropic.com/claude-code](https://www.anthropic.com/claude-code)
 2. Or add Claude to your PATH if it's installed elsewhere
 
@@ -500,6 +541,7 @@ Claude Code reads these variables and uses them instead of the defaults. Simple!
 **Problem**: You didn't source your shell config after installation.
 
 **Solution**: Run the source command the installer showed you:
+
 ```bash
 source ~/.zshrc  # or ~/.bashrc
 ```
@@ -511,11 +553,13 @@ source ~/.zshrc  # or ~/.bashrc
 **Problem**: API key issues.
 
 **Solutions for ccg/ccf/ccg45**:
+
 1. **Check your key**: Visit [z.ai/manage-apikey/apikey-list](https://z.ai/manage-apikey/apikey-list)
 2. **Verify credits**: Make sure your Z.AI account has available credits
 3. **Update the key**: Run `bash install.sh` and choose "Update API key only"
 
 **Solutions for ccx**:
+
 1. **Check your .env file**: Edit `~/.claude-proxy/.env`
 2. **Verify keys are set**: Make sure the API keys for the providers you're using are filled in
 3. **No empty values**: If you're not using a provider, either leave it blank or remove the line
@@ -526,6 +570,7 @@ source ~/.zshrc  # or ~/.bashrc
 **Problem**: Proxy fails to start or times out.
 
 **Solutions**:
+
 1. **Check logs**: Look at `/tmp/claude-proxy.log` (Unix) or `%TEMP%\claude-proxy.log` (Windows)
 2. **Port in use**: Another process might be using port 17870. Set `CLAUDE_PROXY_PORT=17871` in .env
 3. **Missing dependencies**: Run `npm install -g tsx` to ensure TypeScript runner is available
@@ -536,6 +581,7 @@ source ~/.zshrc  # or ~/.bashrc
 **Problem**: `/model` command doesn't seem to work.
 
 **Solutions**:
+
 1. **Check provider prefix**: Use format `/model provider:model-name` (e.g., `/model openai:gpt-4o`)
 2. **Verify API key**: Make sure the provider's API key is set in `~/.claude-proxy/.env`
 3. **Check proxy logs**: Look for errors in `/tmp/claude-proxy.log`
@@ -545,6 +591,7 @@ source ~/.zshrc  # or ~/.bashrc
 **Problem**: Using `ccg` but it's using the wrong API.
 
 **Solution**: Each command is independent. Make sure you:
+
 - Exit any running Claude Code session
 - Start fresh with the command you want (`ccg`, `ccg45`, `ccf`, or `cc`)
 
@@ -555,6 +602,7 @@ source ~/.zshrc  # or ~/.bashrc
 **Problem**: PowerShell execution policy prevents running scripts.
 
 **Solution**:
+
 ```powershell
 Set-ExecutionPolicy -Scope CurrentUser RemoteSigned
 ```
@@ -564,6 +612,7 @@ Set-ExecutionPolicy -Scope CurrentUser RemoteSigned
 **Problem**: PowerShell profile wasn't reloaded after installation.
 
 **Solutions**:
+
 1. Reload profile: `. $PROFILE`
 2. Or restart PowerShell
 3. Or run the full command: `claude-glm`
@@ -585,6 +634,7 @@ Set-ExecutionPolicy -Scope CurrentUser RemoteSigned
 ### macOS / Linux
 
 **Remove wrapper scripts:**
+
 ```bash
 rm ~/.local/bin/claude-glm
 rm ~/.local/bin/claude-glm-4.6
@@ -593,6 +643,7 @@ rm ~/.local/bin/claude-glm-fast
 ```
 
 **Remove config directories** (optional - deletes chat history):
+
 ```bash
 rm -rf ~/.claude-glm
 rm -rf ~/.claude-glm-46
@@ -601,6 +652,7 @@ rm -rf ~/.claude-glm-fast
 ```
 
 **Remove aliases** from `~/.zshrc` or `~/.bashrc`:
+
 ```bash
 # Delete these lines:
 # Claude Code Model Switcher Aliases
@@ -616,6 +668,7 @@ Then run: `source ~/.zshrc`
 ### Windows (PowerShell)
 
 **Remove wrapper scripts:**
+
 ```powershell
 Remove-Item "$env:USERPROFILE\.local\bin\claude-glm.ps1"
 Remove-Item "$env:USERPROFILE\.local\bin\claude-glm-4.6.ps1"
@@ -624,6 +677,7 @@ Remove-Item "$env:USERPROFILE\.local\bin\claude-glm-fast.ps1"
 ```
 
 **Remove config directories** (optional - deletes chat history):
+
 ```powershell
 Remove-Item -Recurse "$env:USERPROFILE\.claude-glm"
 Remove-Item -Recurse "$env:USERPROFILE\.claude-glm-46"
@@ -632,6 +686,7 @@ Remove-Item -Recurse "$env:USERPROFILE\.claude-glm-fast"
 ```
 
 **Remove aliases** from PowerShell profile:
+
 ```powershell
 notepad $PROFILE
 # Delete these lines:
@@ -648,18 +703,24 @@ Then reload: `. $PROFILE`
 ## FAQ
 
 ### Q: Will this affect my existing Claude Code setup?
+
 **A**: No! Your regular Claude Code setup is completely untouched. The wrappers use separate config directories.
 
 ### Q: Can I use both GLM and Claude in the same project?
+
 **A**: Yes! Just use `ccg` for GLM sessions and `cc` for Claude sessions. Each maintains its own chat history. Or use `ccx` to switch between providers in a single session.
 
 ### Q: Which should I use: ccx or dedicated wrappers (ccg/ccf)?
+
 **A**:
+
 - **Use ccx** if you want to switch between multiple providers (OpenAI, Gemini, OpenRouter, GLM, Anthropic) in the same session
 - **Use dedicated wrappers** if you want separate chat histories for different models/providers
 
 ### Q: Which model should I use?
+
 **A**:
+
 - Use **`ccx`** for: Maximum flexibility, model comparison, leveraging different model strengths
 - Use **`ccg` (GLM-4.7)** for: Latest model, complex coding, refactoring, detailed explanations
 - Use **`ccg46` (GLM-4.6)** for: Previous version, if you need consistency with older projects
@@ -668,22 +729,28 @@ Then reload: `. $PROFILE`
 - Use **`cc` (Claude)** for: Your regular Anthropic Claude setup
 
 ### Q: How do I switch models in ccx?
+
 **A**: Use the `/model` command with the format `<provider>:<model-name>`. For example:
+
 - `/model openai:gpt-4o`
 - `/model gemini:gemini-1.5-pro`
 - `/model glm:glm-4.7`
 - `/model glm:glm-4.6`
 
 ### Q: Is this secure?
+
 **A**: Yes! Your API keys are stored locally on your machine in wrapper scripts (bash or PowerShell, depending on your OS). Keep your scripts directory secure with appropriate permissions.
 
 ### Q: Does this work on Windows?
+
 **A**: Yes! Use the PowerShell installer (install.ps1). Windows, macOS, and Linux are all fully supported.
 
 ### Q: Can I use a different Z.AI model?
+
 **A**: Yes! Edit the wrapper scripts in `~/.local/bin/` and change the `ANTHROPIC_MODEL` variable to any model Z.AI supports.
 
 ### Q: What happens if I run out of Z.AI credits?
+
 **A**: The GLM commands will fail with an API error. Just switch to regular Claude using `cc` until you add more credits.
 
 ## Contributing

package/bin/cli.js

@@ -1,9 +1,14 @@
 #!/usr/bin/env node
 
-const { spawn } = require('child_process');
-const path = require('path');
-const os = require('os');
-const fs = require('fs');
+import { spawn } from 'child_process';
+import path from 'path';
+import os from 'os';
+import fs from 'fs';
+import tty from 'tty';
+import { fileURLToPath } from 'url';
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = path.dirname(__filename);
 
 const platform = os.platform();
 const rootDir = path.join(__dirname, '..');
@@ -15,6 +20,9 @@ console.log(`Detected OS: ${platform}\n`);
 function runInstaller() {
   let command, args, scriptPath;
 
+  // Check if running in interactive mode (TTY available)
+  const isInteractive = tty.isatty(0);
+
   if (platform === 'win32') {
     // Windows - run PowerShell installer
     console.log('🪟 Running Windows PowerShell installer...\n');
@@ -51,9 +59,10 @@ function runInstaller() {
     process.exit(1);
   }
 
-  // Spawn the installer process
+  // For interactive mode, we need to pass stdin through properly
+  // Use 'inherit' for all stdio to allow user interaction
   const installer = spawn(command, args, {
-    stdio: 'inherit',
+    stdio: ['inherit', 'inherit', 'inherit'],
     cwd: rootDir
   });
 

package/install.sh

@@ -4,16 +4,23 @@
 #
 # Usage:
 #   Test error reporting:
-#     CLAUDE_GLM_TEST_ERROR=1 bash <(curl -fsSL https://raw.githubusercontent.com/JoeInnsp23/claude-glm-wrapper/main/install.sh)
+#     CLAUDE_GLM_TEST_ERROR=1 bash <(curl -fsSL https://raw.githubusercontent.com/MohMaya/claude-glm-wrapper/main/install.sh)
 #     OR: ./install.sh --test-error
 #
 #   Enable debug mode:
-#     CLAUDE_GLM_DEBUG=1 bash <(curl -fsSL https://raw.githubusercontent.com/JoeInnsp23/claude-glm-wrapper/main/install.sh)
+#     CLAUDE_GLM_DEBUG=1 bash <(curl -fsSL https://raw.githubusercontent.com/MohMaya/claude-glm-wrapper/main/install.sh)
 #     OR: ./install.sh --debug
+#
+#   Non-interactive installation:
+#     ./install.sh --reinstall --api-key=YOUR_KEY
+#     ./install.sh --reinstall --skip-ccx
 
 # Parse command-line arguments
 TEST_ERROR=false
 DEBUG=false
+REINSTALL=false
+API_KEY_ARG=""
+SKIP_CCX=false
 
 for arg in "$@"; do
     case $arg in
@@ -25,6 +32,18 @@ for arg in "$@"; do
             DEBUG=true
             shift
             ;;
+        --reinstall)
+            REINSTALL=true
+            shift
+            ;;
+        --skip-ccx)
+            SKIP_CCX=true
+            shift
+            ;;
+        --api-key=*)
+            API_KEY_ARG="${arg#*=}"
+            shift
+            ;;
         *)
             # Unknown option
             ;;
@@ -40,6 +59,18 @@ if [ "$CLAUDE_GLM_DEBUG" = "1" ] || [ "$CLAUDE_GLM_DEBUG" = "true" ]; then
     DEBUG=true
 fi
 
+if [ "$CLAUDE_GLM_REINSTALL" = "1" ] || [ "$CLAUDE_GLM_REINSTALL" = "true" ]; then
+    REINSTALL=true
+fi
+
+# Detect if running in interactive mode (TTY available)
+INTERACTIVE=true
+if [ ! -t 0 ]; then
+    INTERACTIVE=false
+    # In non-interactive mode, default to reinstall
+    REINSTALL=true
+fi
+
 # Configuration
 USER_BIN_DIR="$HOME/.local/bin"
 GLM_CONFIG_DIR="$HOME/.claude-glm"
@@ -82,15 +113,23 @@ report_error() {
     # Ask if user wants to report the error
     echo "Would you like to report this error to GitHub?"
     echo "This will open your browser with a pre-filled issue report."
-    read -p "Report error? (y/n): " report_choice
+    
+    local report_choice="n"
+    if [ "$INTERACTIVE" = true ]; then
+        read -p "Report error? (y/n): " report_choice
+    else
+        echo "ℹ️  Non-interactive mode: skipping error report prompt"
+    fi
     echo ""
 
     if [ "$report_choice" != "y" ] && [ "$report_choice" != "Y" ]; then
         echo "Error not reported. You can get help at:"
-        echo "  https://github.com/JoeInnsp23/claude-glm-wrapper/issues"
-        echo ""
-        echo "Press Enter to finish..."
-        read
+        echo "  https://github.com/MohMaya/claude-glm-wrapper/issues"
+        if [ "$INTERACTIVE" = true ]; then
+            echo ""
+            echo "Press Enter to finish..."
+            read
+        fi
         return
     fi
 
@@ -151,7 +190,7 @@ $sanitized_error
         encoded_title="Installation%20Error%3A%20Unix%2FLinux%2FmacOS"
     fi
 
-    local issue_url="https://github.com/JoeInnsp23/claude-glm-wrapper/issues/new?title=${encoded_title}&body=${encoded_body}&labels=bug,unix,installation"
+    local issue_url="https://github.com/MohMaya/claude-glm-wrapper/issues/new?title=${encoded_title}&body=${encoded_body}&labels=bug,unix,installation"
 
     echo "📋 Error details have been prepared for reporting."
     echo ""
@@ -185,9 +224,11 @@ $sanitized_error
         echo "After submitting (or if you choose not to), return here."
     fi
 
-    echo ""
-    echo "Press Enter to continue..."
-    read
+    if [ "$INTERACTIVE" = true ]; then
+        echo ""
+        echo "Press Enter to continue..."
+        read
+    fi
 }
 
 # Find all existing wrapper installations
@@ -257,7 +298,15 @@ cleanup_old_wrappers() {
     fi
 
     echo ""
-    read -p "Would you like to clean up old installations? (y/n): " cleanup_choice
+    
+    local cleanup_choice="n"
+    if [ "$INTERACTIVE" = true ]; then
+        read -p "Would you like to clean up old installations? (y/n): " cleanup_choice
+    else
+        # In non-interactive mode, auto-clean old installations
+        cleanup_choice="y"
+        echo "ℹ️  Non-interactive mode: auto-cleaning old installations..."
+    fi
 
     if [[ "$cleanup_choice" == "y" || "$cleanup_choice" == "Y" ]]; then
         echo ""
@@ -800,10 +849,15 @@ check_claude_installation() {
         echo "2. Install Claude Code from: https://www.anthropic.com/claude-code"
         echo "3. Continue anyway (wrappers will be created but won't work until claude is available)"
         echo ""
-        read -p "Continue with installation? (y/n): " continue_choice
-        if [[ "$continue_choice" != "y" && "$continue_choice" != "Y" ]]; then
-            echo "Installation cancelled."
-            exit 1
+        
+        if [ "$INTERACTIVE" = true ]; then
+            read -p "Continue with installation? (y/n): " continue_choice
+            if [[ "$continue_choice" != "y" && "$continue_choice" != "Y" ]]; then
+                echo "Installation cancelled."
+                exit 1
+            fi
+        else
+            echo "ℹ️  Non-interactive mode: continuing anyway..."
         fi
         return 1
     fi
@@ -833,47 +887,66 @@ main() {
     if [ -f "$USER_BIN_DIR/claude-glm" ] || [ -f "$USER_BIN_DIR/claude-glm-fast" ]; then
         echo ""
         echo "✅ Existing installation detected!"
-        echo "1) Update API key only"
-        echo "2) Reinstall everything"
-        echo "3) Cancel"
-        read -p "Choice (1-3): " update_choice
         
-        case "$update_choice" in
-            1)
-                read -p "Enter your Z.AI API key: " input_key
-                if [ -n "$input_key" ]; then
-                    ZAI_API_KEY="$input_key"
-                    create_claude_glm_wrapper
-                    create_claude_glm_46_wrapper
-                    create_claude_glm_45_wrapper
-                    create_claude_glm_fast_wrapper
-                    echo "✅ API key updated!"
+        if [ "$REINSTALL" = true ]; then
+            echo "ℹ️  Reinstalling (--reinstall flag or non-interactive mode)..."
+        elif [ "$INTERACTIVE" = true ]; then
+            echo "1) Update API key only"
+            echo "2) Reinstall everything"
+            echo "3) Cancel"
+            read -p "Choice (1-3): " update_choice
+            
+            case "$update_choice" in
+                1)
+                    read -p "Enter your Z.AI API key: " input_key
+                    if [ -n "$input_key" ]; then
+                        ZAI_API_KEY="$input_key"
+                        create_claude_glm_wrapper
+                        create_claude_glm_46_wrapper
+                        create_claude_glm_45_wrapper
+                        create_claude_glm_fast_wrapper
+                        echo "✅ API key updated!"
+                        exit 0
+                    fi
+                    ;;
+                2)
+                    echo "Reinstalling..."
+                    ;;
+                *)
                     exit 0
-                fi
-                ;;
-            2)
-                echo "Reinstalling..."
-                ;;
-            *)
-                exit 0
-                ;;
-        esac
+                    ;;
+            esac
+        else
+            echo "ℹ️  Non-interactive mode: reinstalling..."
+        fi
     fi
     
-    # Get API key
-    echo ""
-    echo "Enter your Z.AI API key (from https://z.ai/manage-apikey/apikey-list)"
-    read -p "API Key: " input_key
-    
-    if [ -n "$input_key" ]; then
-        ZAI_API_KEY="$input_key"
-        echo "✅ API key received (${#input_key} characters)"
+    # Get API key (from argument, environment, or prompt)
+    if [ -n "$API_KEY_ARG" ]; then
+        ZAI_API_KEY="$API_KEY_ARG"
+        echo "✅ API key provided via argument (${#ZAI_API_KEY} characters)"
+    elif [ -n "${ZAI_API_KEY_ENV:-}" ]; then
+        ZAI_API_KEY="$ZAI_API_KEY_ENV"
+        echo "✅ API key provided via environment (${#ZAI_API_KEY} characters)"
+    elif [ "$INTERACTIVE" = true ]; then
+        echo ""
+        echo "Enter your Z.AI API key (from https://z.ai/manage-apikey/apikey-list)"
+        read -p "API Key: " input_key
+        
+        if [ -n "$input_key" ]; then
+            ZAI_API_KEY="$input_key"
+            echo "✅ API key received (${#input_key} characters)"
+        else
+            echo "⚠️  No API key provided. Add it manually later to:"
+            echo "   $USER_BIN_DIR/claude-glm"
+            echo "   $USER_BIN_DIR/claude-glm-4.6"
+            echo "   $USER_BIN_DIR/claude-glm-4.5"
+            echo "   $USER_BIN_DIR/claude-glm-fast"
+        fi
     else
-        echo "⚠️  No API key provided. Add it manually later to:"
-        echo "   $USER_BIN_DIR/claude-glm"
-        echo "   $USER_BIN_DIR/claude-glm-4.6"
-        echo "   $USER_BIN_DIR/claude-glm-4.5"
-        echo "   $USER_BIN_DIR/claude-glm-fast"
+        echo "⚠️  Non-interactive mode: No API key provided."
+        echo "   Use --api-key=YOUR_KEY or set ZAI_API_KEY_ENV environment variable."
+        echo "   Or add it manually later to the wrapper scripts."
     fi
     
     # Create wrappers
@@ -884,17 +957,26 @@ main() {
     create_shell_aliases
 
     # Ask about ccx installation
-    echo ""
-    echo "📦 Multi-Provider Proxy (ccx)"
-    echo "================================"
-    echo "ccx allows you to switch between multiple AI providers in a single session:"
-    echo "  • OpenAI (GPT-4, GPT-4o, etc.)"
-    echo "  • OpenRouter (access to many models)"
-    echo "  • Google Gemini"
-    echo "  • Z.AI GLM models"
-    echo "  • Anthropic Claude"
-    echo ""
-    read -p "Install ccx? (Y/n): " install_ccx_choice
+    local install_ccx_choice="Y"
+    
+    if [ "$SKIP_CCX" = true ]; then
+        install_ccx_choice="n"
+        echo "ℹ️  Skipping ccx installation (--skip-ccx flag)"
+    elif [ "$INTERACTIVE" = true ]; then
+        echo ""
+        echo "📦 Multi-Provider Proxy (ccx)"
+        echo "================================"
+        echo "ccx allows you to switch between multiple AI providers in a single session:"
+        echo "  • OpenAI (GPT-4, GPT-4o, etc.)"
+        echo "  • OpenRouter (access to many models)"
+        echo "  • Google Gemini"
+        echo "  • Z.AI GLM models"
+        echo "  • Anthropic Claude"
+        echo ""
+        read -p "Install ccx? (Y/n): " install_ccx_choice
+    else
+        echo "ℹ️  Non-interactive mode: installing ccx by default (use --skip-ccx to skip)"
+    fi
 
     if [ "$install_ccx_choice" != "n" ] && [ "$install_ccx_choice" != "N" ]; then
         install_ccx
@@ -970,8 +1052,10 @@ handle_error() {
     # Give user time to read any final messages before stopping
     echo ""
     echo "Installation terminated due to error."
-    echo "Press Enter to finish (window will remain open)..."
-    read
+    if [ "$INTERACTIVE" = true ]; then
+        echo "Press Enter to finish (window will remain open)..."
+        read
+    fi
     # Return to stop script execution without closing terminal
     return
 }
@@ -996,10 +1080,12 @@ if [ "$TEST_ERROR" = true ]; then
     echo "✅ Test complete. If a browser window opened, error reporting is working!"
     echo ""
     echo "To run normal installation, use:"
-    echo "   curl -fsSL https://raw.githubusercontent.com/JoeInnsp23/claude-glm-wrapper/main/install.sh | bash"
+    echo "   curl -fsSL https://raw.githubusercontent.com/MohMaya/claude-glm-wrapper/main/install.sh | bash"
     echo ""
-    echo "Press Enter to finish (window will remain open)..."
-    read
+    if [ "$INTERACTIVE" = true ]; then
+        echo "Press Enter to finish (window will remain open)..."
+        read
+    fi
     # Script ends naturally here - terminal stays open
     exit 0
 fi

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-glm-alt-installer",
-  "version": "2.0.0",
+  "version": "2.0.2",
   "description": "Cross-platform installer for Claude Code with Z.AI GLM models and multi-provider proxy. Run with: npx claude-glm-alt-installer",
   "keywords": [
     "claude",