Kyber: run your fleet of AI agents on Kubernetes

Run yourClaude Codeagenton Google Cloud

Long-lived AI agents with their own durable environment, managed as a fleet on Kubernetes and reachable from wherever you work.

Get startedView on GitHub

Pause examples

# Install from the published Helm chart. Nothing to pin, no fork.
helm install kyber oci://ghcr.io/matty-v/charts/kyber \
  --namespace kyber-system --wait
# 15 minutes from an empty cluster to a live fleet console

What you can do

Long-lived, sandboxed agents

Each agent owns a whole pod backed by durable storage. Installed packages, cloned repos, credentials, and memory survive pod restarts and upgrades.

The Kubernetes cluster you choose

Run on Kubernetes 1.33+ with containerd 2.0+: a Windows laptop, a spare box, a Mac, or a cloud cluster. Kyber can provision GCP VMs, spot instances included.

Chat from your phone

Two-way Telegram and Discord conversations with any agent in the fleet. Assign work, approve decisions, and get results without opening a terminal.

Own the whole environment

Choose each agent's machine, VM type, disk, CPU, and memory. Reboot it, stop it, or open a shell into the running agent straight from the console.

Schedules and handoffs

Cron inside each agent survives restarts. Running agents send each other signed messages, so one agent's output becomes another's next prompt.

Your subscription, not per-token

Claude Code agents sign in with your Claude subscription, Codex agents with your ChatGPT subscription. API keys work too, but you do not need them.

How it works

Web consoleAPITelegram + Discord

commands + status

Kubernetes cluster

Kyber control planeAPI · lifecycle controllers · scheduling · telemetry

reconciles

Agent podClaude CodePersistent disk

Agent podCodexPersistent disk

Agent podYour runtimePersistent disk

Use cases

A fully autonomous dev team

A team of agents that takes GitHub issues to reviewed, merged pull requests: planner, builders, and reviewers, each in its own pod.

Read more →

A personal agent fleet

Long-lived assistants that keep their tools, repos, and memory. Always on, on hardware you control.

Read more →

Scheduled AI automations

Agents on cron: nightly audits, weekly reports, monitors that investigate before they page you. The schedules survive restarts.

Read more →

Give your agents somewhere to live

Fifteen minutes from an empty cluster, or a k3d one-liner, to a fleet console with a live agent. No cloud account required.

Read the quickstartStar on GitHub