🚀 Claude Code Router: Master Setup Guide (WSL + Gemini)

📺 Watch the Video: Click here to watch the full step-by-step tutorial

This repository is the official guide for my YouTube tutorial. It helps you set up Claude Code on Windows using WSL2 and powers it with Google Gemini using the Claude Code Router.

---

🛠 Step 0: Install Windows Subsystem for Linux (WSL)

Important: This is the first thing you need to do. We are setting up a Linux environment where Claude Code will run.

1. Open PowerShell as Administrator

• Press the Windows key
• Search for PowerShell
• Click Run as administrator
• Click Yes

2. Run the Install Command

wsl --install

3. Downloading & Launching

• Windows will download and install WSL
• Ubuntu will launch automatically
• Stay in the same window while it provisions

4. Set Your Username & Password

• Enter a username
• Enter a password
• Confirm the password

⚠️ Password typing is hidden. This is normal in Linux.

5. Success

When you see something like:

hammad@PC:/mnt/c/WINDOWS/system32$

Keep this terminal open and continue.

---

⚡ Method 1: Automated "Speedrun" Script

1. Run the Installer

Inside your WSL / Ubuntu terminal, run:

curl -sSL https://raw.githubusercontent.com/devhammad0/claude-code-router-setup/main/scripts/install.sh | bash

After the script finishes:

source ~/.bashrc

2. Set Your Google API Key

Replace YOUR_KEY_HERE with your key from Google AI Studio:

echo 'export GOOGLE_API_KEY="YOUR_KEY_HERE"' >> ~/.bashrc

3. Reload your Environment

source ~/.bashrc

4. Laucnh Claude Code Router

ccr code

---

🛠 Method 2: Manual Installation (Step-by-Step)

1. Update Linux Packages

sudo apt update && sudo apt upgrade -y
sudo apt install -y curl git build-essential

2. Install Node.js (via NVM)

curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.1/install.sh | bash
source ~/.bashrc
nvm install --lts

3. Install Claude Code & CCR Router

npm install -g @anthropic-ai/claude-code @musistudio/claude-code-router

---

⚙️ Configuration Setup

Step 1: Check Your Shell (Important)

Run:

echo $SHELL

• If output contains /bash → use ~/.bashrc
• If output contains /zsh → use ~/.zshrc

---

Step 2: Create Config File (Google Gemini)

mkdir -p ~/.claude-code-router ~/.claude

Run the following command to create your configuration:

cat > ~/.claude-code-router/config.json << 'EOF'
{
  "LOG": true,
  "LOG_LEVEL": "info",
  "HOST": "127.0.0.1",
  "PORT": 3456,
  "APIKEY": "your_secure_password_here",
  "API_TIMEOUT_MS": 600000,
  "Providers": [
    {
      "name": "gemini",
      "api_base_url": "https://generativelanguage.googleapis.com/v1beta/models/",
      "api_key": "$GOOGLE_API_KEY",
      "models": [
        "gemini-3-flash-preview",
        "gemini-3-pro-preview",
        "gemini-2.5-flash-lite",
        "gemini-2.0-flash"
      ],
      "transformer": { "use": ["gemini"] }
    }
  ],
  "Router": {
    "default": "gemini,gemini-2.5-flash-lite",
    "background": "gemini,gemini-2.5-flash-lite",
    "think": "gemini,gemini-2.5-flash-lite",
    "longContext": "gemini,gemini-2.5-flash-lite",
    "longContextThreshold": 60000
  }
}
EOF

---

📊 Model Performance & Quota (Free Tier)

Before you run the agent, check the current limits for the models included in the config above. These limits are reset at Midnight Pacific Time (PT).

Model Name	RPM (Req/Min)	TPM (Tokens/Min)	RPD (Req/Day)	Status/Notes
gemini-2.5-flash-lite	30	1,000,000	1,500	✅ Recommended for Coding
gemini-2.0-flash	10	1,000,000	1,000	Stable legacy option
gemini-3-flash-preview	2	100,000	100	⚠️ Strict preview limits
gemini-3-pro-preview	0 - 2*	32,000	5 - 10	🚫 Effectively Paid-Only

---

Free Tier Quota Note:

• RPM (Requests Per Minute) is strictly enforced. If Claude Code fires multiple background file reads, you may hit a brief 429 error.
• Gemini 3 Pro Preview is currently highly restricted for free users; frequent use in the think or longContext routes will likely exhaust your daily quota (RPD) in minutes.

TIP:
*Why use Flash-Lite? Claude Code performs many small background tasks (reading files, checking status). Using a model with less than 10 RPM will cause "Rate Limit" errors almost immediately. Flash-Lite is the only free model that provides a smooth experience for complex coding.

---

Step 3: Set API Key

Use the shell configuration file you identified in Step 1. Replace YOUR_KEY_HERE with your actual Google AI Studio API key.

If using bash:

echo 'export GOOGLE_API_KEY="YOUR_KEY_HERE"' >> ~/.bashrc
source ~/.bashrc

If using zsh:

echo 'export GOOGLE_API_KEY="YOUR_KEY_HERE"' >> ~/.zshrc
source ~/.zshrc

---

🚀 How to Run

ccr code

---

💡 Essential Commands

• ccr start / ccr stop / ccr status — Manage the background router service.
• ccr model — Quick interactive configuration of your model providers.
• ccr ui → Dashboard at http://localhost:3456

---

❓ Troubleshooting

1. Service Startup Timeout

If ccr code shows a "Timeout" error, the background service might just need a manual kickstart. Run these separately:

ccr start
ccr code

2. Status shows "Not Running"

This usually means a "ghost" process is already using the port. Run this to clear it:

# Find any process on port 3456
sudo lsof -i :3456
# Use the PID found to kill it: sudo kill -9 <PID>

3. Verification Commands

Use these commands to verify your setup:

• ccr status — Check if the background service is running.
• /status — (Inside Claude) Verify the Base URL is 127.0.0.1:3456.

---

Maintained by Muhammad Hammad | LinkedIn Profile