Reference
| Variable | Required | Default | Description |
|---|
ANTHROPIC_API_KEY | At least one API key | — | API key for Anthropic (Claude) models. Get one at console.anthropic.com. |
OPENAI_API_KEY | At least one API key | — | API key for OpenAI (GPT-4o) models. Get one at platform.openai.com. |
GHIDRA_INSTALL_DIR | No | Auto-detected | Path to your Ghidra installation directory. Only needed if auto-detection fails. |
JAVA_HOME | No | Auto-detected | Path to your JDK installation. Only needed if auto-detection fails. |
KONG_CONFIG_DIR | No | ~/.config/kong | Override the directory where Kong stores its configuration database. |
API keys
You need at least one API key — either Anthropic or OpenAI. If you’re using a custom endpoint, no API key is required (or you can set one during kong setup if your endpoint requires authentication).
# Set your API key(s)
export ANTHROPIC_API_KEY="sk-ant-api03-..."
export OPENAI_API_KEY="sk-..."
Add these to your shell profile (~/.zshrc, ~/.bashrc) so they persist across sessions.
Ghidra and JDK auto-detection
Kong searches common installation paths for Ghidra and JDK. You only need to set GHIDRA_INSTALL_DIR and JAVA_HOME if:
- You installed Ghidra in a non-standard location
- Auto-detection picks the wrong version
- You’re running in a container or CI environment
export GHIDRA_INSTALL_DIR="/opt/ghidra_11.3"
export JAVA_HOME="/usr/lib/jvm/java-21-openjdk"
Config directory
Kong stores its configuration (enabled providers, default provider, custom endpoint settings) in an SQLite database at ~/.config/kong/config.db. Override the location:
export KONG_CONFIG_DIR="/path/to/custom/config"
Further reading
