Installation
Kairos Substrate is distributed as a native package for Linux and macOS. The installer places only the kairos binary on your system PATH — no configuration files, services, or databases are created.
Native Package Install
Section titled “Native Package Install”Linux (.deb)
Section titled “Linux (.deb)”Download kairos-substrate_<version>_amd64.deb from the release artifacts, then install:
sudo dpkg -i kairos-substrate_<version>_amd64.debVerify:
kairos --helpUninstall
Section titled “Uninstall”sudo dpkg -r kairos-substratemacOS (.pkg)
Section titled “macOS (.pkg)”Download kairos-substrate-<version>-arm64.pkg from the release artifacts.
The package installs the binary to /usr/local/bin/kairos.
Verify:
kairos --helpUninstall
Section titled “Uninstall”macOS packages do not include a native uninstall mechanism. Remove the binary manually:
sudo rm /usr/local/bin/kairosTo also remove the package receipt (optional):
sudo pkgutil --forget dev.anankelabs.kairos-substrateVerify Download Integrity
Section titled “Verify Download Integrity”Each release includes a SHA256SUMS file listing all release artifacts. To verify a specific download, extract its line from the manifest:
grep kairos-substrate_<version>_amd64.deb SHA256SUMS | sha256sum -c -Starter Kit
Section titled “Starter Kit”The starter kit is a separate versioned archive containing reference configuration files and quick-start documentation. It is not required for CLI usage.
Download kairos-substrate-starter-kit-<version>.tar.gz from the release artifacts:
tar xzf kairos-substrate-starter-kit-<version>.tar.gzContents:
| File | Purpose |
|---|---|
README.md | Quick-start guide |
request.example.json | Example evaluation request |
artifact.example.json | Example artifact placeholder |
policy.example.json | Example policy placeholder |
These files are illustrative. Use the signed deployment files provided for your licensed environment when running real evaluations.
Next Steps
Section titled “Next Steps”After installing, you need to activate a license before running evaluations.