Files

Bytebot Helm Charts

This directory contains Helm charts for deploying Bytebot on Kubernetes.

Structure

helm/
├── Chart.yaml              # Main chart definition
├── values.yaml             # Default configuration values
├── templates/              # Kubernetes manifests templates
│   ├── NOTES.txt          # Post-installation notes
│   └── ingress.yaml       # Optional ingress configuration
└── charts/                 # Individual service charts
    ├── bytebot-desktop/    # Desktop VNC service
    ├── bytebot-agent/      # Backend API service
    ├── bytebot-ui/         # Frontend UI service
    └── postgresql/         # PostgreSQL database

Prerequisites

  • Kubernetes 1.19+
  • Helm 3.0+
  • PV provisioner support (for PostgreSQL persistence)

Installation

  1. Add required Helm repositories (for PostgreSQL dependency):

    helm repo add bitnami https://charts.bitnami.com/bitnami
    helm repo update
    
  2. Standard Installation (API keys on agent):

    cd helm
    helm dependency update
    helm install bytebot . --namespace bytebot --create-namespace \
      --set bytebot-agent.env.ANTHROPIC_API_KEY="your-api-key" \
      --set bytebot-agent.env.OPENAI_API_KEY="your-api-key" \
      --set bytebot-agent.env.GEMINI_API_KEY="your-api-key"
    
  3. Installation with LiteLLM Proxy (recommended for production):

    cd helm
    helm dependency update
    helm install bytebot . -f values-proxy.yaml \
      --namespace bytebot \
      --create-namespace \
      --set bytebot-llm-proxy.env.ANTHROPIC_API_KEY="your-api-key" \
      --set bytebot-llm-proxy.env.OPENAI_API_KEY="your-api-key" \
      --set bytebot-llm-proxy.env.GEMINI_API_KEY="your-api-key"
    

    The proxy setup provides:

    • Centralized API key management
    • Request routing and load balancing
    • Rate limiting and retry logic
    • Support for multiple LLM providers

Configuration

Key Configuration Options

Parameter Description Default
postgresql.enabled Enable PostgreSQL subchart true
postgresql.auth.password PostgreSQL password postgres
postgresql.auth.database PostgreSQL database name bytebotdb
bytebot-desktop.image.tag Desktop service image tag edge
bytebot-agent.env.ANTHROPIC_API_KEY Anthropic API key (standard mode) ""
bytebot-agent.env.OPENAI_API_KEY OpenAI API key (standard mode) ""
bytebot-agent.env.GEMINI_API_KEY Gemini API key (standard mode) ""
bytebot-llm-proxy.enabled Enable LiteLLM proxy false
bytebot-llm-proxy.env.ANTHROPIC_API_KEY Anthropic API key (proxy mode) ""
bytebot-llm-proxy.env.OPENAI_API_KEY OpenAI API key (proxy mode) ""
bytebot-llm-proxy.env.GEMINI_API_KEY Gemini API key (proxy mode) ""
bytebot-ui.service.type UI service type ClusterIP
ingress.enabled Enable ingress false
ingress.host Ingress hostname bytebot.example.com

Using External PostgreSQL

To use an external PostgreSQL database instead of the included one:

postgresql:
  enabled: false

bytebot-agent:
  externalDatabase:
    host: "your-postgres-host"
    port: 5432
    database: "bytebotdb"
    username: "postgres"
    password: "your-password"

Exposing Services

Using LoadBalancer:

bytebot-ui:
  service:
    type: LoadBalancer

Using Ingress:

ingress:
  enabled: true
  className: "nginx"
  host: bytebot.yourdomain.com
  tls:
    enabled: true
    secretName: bytebot-tls
  annotations:
    cert-manager.io/cluster-issuer: "letsencrypt-prod"

Accessing Bytebot

After installation, follow the notes displayed by Helm or use:

helm status bytebot -n bytebot

Port Forwarding (for local access):

# All services at once
kubectl port-forward -n bytebot svc/bytebot-ui 9992:9992 &
kubectl port-forward -n bytebot svc/bytebot-agent 9991:9991 &
kubectl port-forward -n bytebot svc/bytebot-desktop 9990:9990 &

# If using LiteLLM proxy, also forward:
kubectl port-forward -n bytebot svc/bytebot-llm-proxy 4000:4000 &

Then access:

Upgrading

helm upgrade bytebot . -n bytebot

Uninstalling

helm uninstall bytebot -n bytebot

Troubleshooting

  1. Check pod status:

    kubectl get pods -n bytebot
    
  2. View logs:

    kubectl logs -n bytebot deployment/bytebot-agent
    
  3. Describe resources:

    kubectl describe pod -n bytebot <pod-name>
    

Deployment Modes

Standard Mode

In standard mode, API keys are configured directly on the agent service. This is simpler but requires updating the agent deployment to change API keys.

In proxy mode, the LiteLLM proxy handles all LLM API calls. Benefits include:

  • Centralized API key management
  • Easy switching between LLM providers
  • Built-in rate limiting and retry logic
  • Request logging and monitoring capabilities

To switch from standard to proxy mode:

helm upgrade bytebot . -f values-proxy.yaml -n bytebot \
  --set bytebot-llm-proxy.env.ANTHROPIC_API_KEY="your-key" \
  --set bytebot-llm-proxy.env.OPENAI_API_KEY="your-key" \
  --set bytebot-llm-proxy.env.GEMINI_API_KEY="your-key"

Individual Chart Usage

Each service can also be deployed independently:

# Deploy only the desktop service
helm install bytebot-desktop ./charts/bytebot-desktop

# Deploy only the agent service
helm install bytebot-agent ./charts/bytebot-agent

# Deploy only the LiteLLM proxy
helm install bytebot-llm-proxy ./charts/bytebot-llm-proxy