macOS / Linux

curl -fsSL ccs.kaitran.ca/install | bash

Windows PowerShell

irm ccs.kaitran.ca/install | iex

Note: Traditional installs bypass Node.js routing for faster startup, but npm is prioritized for easier deployment automation.

If Claude CLI is installed in a non-standard location (D drive, custom directory), set CCS_CLAUDE_PATH:

export CCS_CLAUDE_PATH="/path/to/claude"              # Unix
$env:CCS_CLAUDE_PATH = "D:\Tools\Claude\claude.exe"   # Windows

See also: Troubleshooting Guide for detailed setup instructions.

Windows users: Enable Developer Mode for true symlinks (better performance, instant sync):

1. Open Settings → Privacy & Security → For developers
2. Enable Developer Mode
3. Reinstall CCS: npm install -g @kaitranntt/ccs

Warning: Without Developer Mode, CCS automatically falls back to copying directories (works but no instant sync across profiles).

# Kimi - Stable thinking support
ccs kimi "Design a caching strategy with trade-off analysis"

# GLMT - Experimental (see full disclaimer below)
ccs glmt "Debug complex algorithm with reasoning steps"

Note: GLMT is experimental and unstable. See GLM with Thinking (GLMT) section below for full details.

graph LR
    A[2pm: Building features<br/>In the zone] --> B[3pm: Usage limit hit<br/>BLOCKED]
    B --> C[3:05pm: Stop work<br/>Edit settings.json]
    C --> D[3:15pm: Switch accounts<br/>Context lost]
    D --> E[3:30pm: Restart<br/>Trying to focus]
    E --> F[4pm: Finally productive<br/>Back in flow]

    style A fill:#d4edda,stroke:#333,color:#000
    style B fill:#f8d7da,stroke:#333,color:#000
    style C fill:#fff3cd,stroke:#333,color:#000
    style D fill:#f8d7da,stroke:#333,color:#000
    style E fill:#fff3cd,stroke:#333,color:#000
    style F fill:#d4edda,stroke:#333,color:#000

Result: 1 hour lost, momentum destroyed, frustration builds

graph LR
    A[2pm: Start work] --> B[Terminal 1: Claude Pro<br/>Strategic planning]
    A --> C[Terminal 2: GLM<br/>Code execution]
    B --> D[3pm: Still shipping<br/>No interruptions]
    C --> D
    D --> E[4pm: Flow state<br/>Productivity peak]
    E --> F[5pm: Features shipped<br/>Context maintained]

    style A fill:#e7f3ff,stroke:#333,color:#000
    style B fill:#cfe2ff,stroke:#333,color:#000
    style C fill:#cfe2ff,stroke:#333,color:#000
    style D fill:#d4edda,stroke:#333,color:#000
    style E fill:#d4edda,stroke:#333,color:#000
    style F fill:#d4edda,stroke:#333,color:#000

Result: Zero downtime, continuous productivity, less frustration

• Setup: Your existing Claude Pro + GLM Lite (cost-effective add-on)
• Value: Save 1 hour/day × 20 workdays = 20 hours/month recovered
• ROI: Your development time is worth more than the setup cost
• Reality: Shipping faster than the overhead

• Best for: Cost-conscious development, basic code generation
• Usage: Just use ccs glm directly for cost-effective AI assistance
• Reality: No Claude access, but capable for many coding tasks
• Setup: GLM API key only, very affordable

• Best for: Daily code delivery, serious development work
• Usage: ccs for planning + ccs glm for execution (parallel workflow)
• Reality: Perfect balance of capability and cost for most developers
• Value: Never hit session limits, continuous productivity

• Best for: Heavy workloads, concurrent projects, solo dev
• Unlocks: Never drain session or weekly limits
• Workflow: 3+ terminals running specialized tasks simultaneously

• When needed: Strict separation of work and personal AI contexts
• Setup: ccs auth create work + ccs auth create personal
• Note: Advanced feature - most users don't need this

GLMT attempts MCP tools and function calling:

• Bidirectional Transformation: Anthropic tools ↔ OpenAI format (unstable)
• MCP Integration: MCP tools sometimes execute (often output XML garbage)
• Streaming Tool Calls: Real-time tool calls (when not crashing)
• Backward Compatible: May break existing thinking support
• Configuration Required: Frequent manual debugging needed

GLMT attempts real-time streaming with incremental reasoning content delivery:

• Default: Streaming enabled (TTFB <500ms when it works)
• Auto-fallback: Frequently switches to buffered mode due to errors
• Thinking parameter: Claude CLI thinking parameter sometimes works
  • May ignore thinking.type and budget_tokens
  • Precedence: CLI parameter > message tags > default (when not broken)

Status: Z.AI (tested, tool calls frequently break, requires constant debugging)

1. CCS spawns embedded HTTP proxy on localhost (if not crashing)
2. Proxy attempts to convert Anthropic format → OpenAI format (often fails)
3. Tries to transform Anthropic tools → OpenAI function calling format (buggy)
4. Forwards to Z.AI with reasoning parameters and tools (when not timing out)
5. Attempts to convert reasoning_content → thinking blocks (partial or broken)
6. Attempts to convert OpenAI tool_calls → Anthropic tool_use blocks (XML garbage common)
7. Thinking and tool calls sometimes appear in Claude Code UI (when not broken)

Control Tags:

• <Thinking:On|Off> - Enable/disable reasoning blocks (default: On)
• <Effort:Low|Medium|High> - Control reasoning depth (deprecated - Z.AI only supports binary thinking)

Thinking Keywords (inconsistent activation):

• think - Sometimes enables reasoning (low effort)
• think hard - Sometimes enables reasoning (medium effort)
• think harder - Sometimes enables reasoning (high effort)
• ultrathink - Attempts maximum reasoning depth (often breaks)

GLMT features (all experimental):

• Forced English output enforcement (sometimes works)
• Random thinking mode activation (unpredictable)
• Attempted streaming with frequent fallback to buffered mode

General:

• CCS_DEBUG_LOG=1 - Enable debug file logging
• CCS_CLAUDE_PATH=/path/to/claude - Custom Claude CLI path

# Edit GLMT settings
nano ~/.ccs/glmt.settings.json

Set Z.AI API key (requires coding plan):

{
  "env": {
    "ANTHROPIC_AUTH_TOKEN": "your-z-ai-api-key"
  }
}

v3.4 Protection Limits:

Enable verbose logging:

ccs glmt --verbose "your prompt"

Enable debug file logging:

export CCS_DEBUG_LOG=1
ccs glmt --verbose "your prompt"
# Logs: ~/.ccs/logs/

GLMT debugging:

# Verbose logging shows streaming status and reasoning details
ccs glmt --verbose "test"

Check reasoning content:

cat ~/.ccs/logs/*response-openai.json | jq '.choices[0].message.reasoning_content'

Troubleshooting:

• If absent: Z.AI API issue (verify key, account status)
• If present: Transformation issue (check response-anthropic.json)