Do I need Kubernetes to use rbee?

No. rbee uses SSH for deployment. No Kubernetes, no Docker, no YAML. If you can SSH into a machine, you can deploy a worker. 5-minute setup vs 6-month Kubernetes project.

How does SSH deployment work?

rbee connects to your machines via SSH, installs the worker, detects hardware, and starts the worker. One command: rbee worker deploy --host gpu-server.local. No Docker containers, no Kubernetes pods.

Can I deploy to bare metal servers?

Yes. rbee works on bare metal, VMs, or cloud instances. As long as you have SSH access, you can deploy a worker. No Docker or Kubernetes required.

How long does deployment take?

5 minutes for SSH deployment vs 6-month Kubernetes project. rbee handles installation, hardware detection, and worker startup automatically. No YAML files, no cluster management.

Is SSH deployment secure?

Yes. rbee uses standard SSH authentication (keys or passwords). Same security as any SSH-based deployment. No additional attack surface from Docker or Kubernetes.

Core Feature • Free Forever

Stop FightingKubernetes.

Deploy with just SSH. No Docker, no Kubernetes, no YAML hell. Works on bare metal, VMs, or cloud. 5-minute setup vs 6-month K8s project. If you can SSH into a machine, you can deploy a worker.

Just SSHNo DockerNo Kubernetes5-minute setup

Try rbee Now View Docs

terminal

$ rbee worker deploy --host gpu-server.local

▊ Streaming tokens...

# Deploy worker via SSH (30 seconds)

✓ Connecting to gpu-server.local via SSH...

✓ Installing rbee worker...

✓ Detecting hardware (NVIDIA RTX 4090)...

✓ Starting worker...

Worker deployed successfully!

# No Docker. No Kubernetes. Just SSH.

GPU 1: 0%

GPU 2: 0%

Cost: 5 min

SSH-based deployment to multiple machines without Kubernetes

Just SSHNo K8s

The Kubernetes Nightmare

Most AI tools force you into Kubernetes. 6-month setup. YAML hell. Operational nightmare.

6-Month Kubernetes Setup

vLLM requires Kubernetes. Ray+KServe takes 6 months to set up. You need a dedicated DevOps team. Most companies give up.

Time waste

YAML Hell

Kubernetes means endless YAML files. ConfigMaps, Deployments, Services, Ingress. One typo breaks everything. Debugging is a nightmare.

Complexity

Operational Overhead

Kubernetes requires constant maintenance. Updates, security patches, cluster management. You spend more time on ops than AI.

High cost

Deploy with Just SSH.

rbee uses SSH for deployment. No Docker, no Kubernetes, no complexity. 5-minute setup.

Just SSH Access

Deploy to any machine with SSH access. No Docker, no Kubernetes, no complex setup. Works on bare metal, VMs, or cloud.

5-Minute Setup

Deploy a worker in 5 minutes vs 6-month Kubernetes project. No YAML files. No cluster management. Just SSH.

Simple Troubleshooting

SSH into the machine and debug. No Kubernetes logs. No pod restarts. Traditional ops tools work.

Lower Costs

No Kubernetes cluster overhead. No dedicated DevOps team. Lower operational costs. Free forever.

How It Works

Three steps to deploy a worker via SSH

Add SSH host

Configure SSH access

# rbee.yaml - SSH configuration
hives:
  - name: gpu-server
    host: gpu-server.local
    ssh:
      user: ubuntu
      key: ~/.ssh/id_rsa

Deploy worker

One command deployment

# Deploy worker via SSH
rbee worker deploy --host gpu-server.local

# rbee handles everything:
# - SSH connection
# - Worker installation
# - Hardware detection
# - Worker startup

Done. No Kubernetes.

Note

Worker is deployed and running. No Docker. No Kubernetes. No YAML. Just SSH. 5-minute setup vs 6-month K8s project.

Deployment Examples

See how easy SSH deployment is

bash

# Deploy worker to remote machine
rbee worker deploy --host gpu-server.local --name worker-1

# Output:
# ✓ Connecting to gpu-server.local via SSH...
# ✓ Installing rbee worker...
# ✓ Detecting hardware (NVIDIA RTX 4090)...
# ✓ Starting worker...
# Worker deployed successfully!

# That's it! No Docker. No Kubernetes.

Use Cases

Who benefits from SSH deployment

Homelabbers

Scenario

Deploy to home machines without Docker or Kubernetes

Solution

rbee SSH deployment (free forever)

Outcome

5-minute setup. No K8s complexity. Traditional ops tools work.

Homelab GPU orchestration

Startups

Scenario

Fast deployment without dedicated DevOps team

Solution

SSH deployment instead of 6-month K8s project

Outcome

Ship AI features in days, not months. No DevOps team needed.

Startup AI deployment

Enterprises

Scenario

Deploy to existing infrastructure without re-architecture

Solution

SSH into existing servers, deploy workers

Outcome

No Kubernetes migration. Use existing infrastructure. Lower costs.

Enterprise deployment

DevOps Teams

Scenario

Simple deployment without Kubernetes complexity

Solution

Traditional SSH-based deployment

Outcome

Familiar tools. Simple troubleshooting. Lower operational overhead.

DevOps-friendly deployment