KCode Installation

Download

Download the latest binary for your platform from kulvex.ai/downloads/kcode:

Quick Install (Linux / macOS)

# Download and install in one command
curl -fsSL https://kulvex.ai/downloads/kcode/kcode-1.1.0-linux-x64 -o kcode
chmod +x kcode
sudo mv kcode /usr/local/bin/
 
# Or install to user directory (no sudo)
mv kcode ~/.local/bin/

For Apple Silicon Macs:

curl -fsSL https://kulvex.ai/downloads/kcode/kcode-1.1.0-macos-arm64 -o kcode
chmod +x kcode
mv kcode ~/.local/bin/

Windows

Download kcode-1.1.0-windows-x64.exe, rename it to kcode.exe, and add it to your PATH.

First Run

# Start interactive mode
kcode
 
# Or single-shot
kcode "hello"

On first run, KCode will:

1. Create ~/.kcode/ config directory
2. Look for a model at http://localhost:10091 (default llama-server)
3. If no model is running, suggest running kcode setup or configuring a provider

Activating Your License

kcode activate <your-license-key>

Your license key looks like klx_lic_xxxxxxxxxxxx. It was emailed to you after purchase.

To check license status:

kcode license

Shell Aliases

Add to your ~/.bashrc or ~/.zshrc:

alias kk="kcode"
alias k="kcode --print"

Or if KULVEX is installed, source the shell init:

source ~/.kulvex/data/kulvex-shell-init.sh

Updating

Download the new binary and replace the old one:

curl -fsSL https://kulvex.ai/downloads/kcode/kcode-1.1.0-linux-x64 -o /tmp/kcode
chmod +x /tmp/kcode
sudo mv /tmp/kcode /usr/local/bin/kcode

Or use the built-in update check:

kcode update

System Requirements

• OS: Linux (x64/ARM64), macOS (Intel/Apple Silicon), Windows (x64)
• RAM: 4 GB minimum (KCode itself uses ~200 MB)
• Disk: ~120 MB for the binary
• For local models: GPU with 8-48+ GB VRAM (see Models & Providers)
• For cloud models: Internet connection + API key