Deployment Options

Horizons OmniChat offers flexible deployment options to meet different organizational needs and requirements.

Available Deployment Modes

1. Local Mode

• Best for: Development, testing, and privacy-focused deployments
• Key Features:
  • Complete local deployment
  • No cloud dependencies
  • Full data privacy
  • Local model inference via Ollama
• Learn More: Local Deployment Guide

2. Hybrid Mode

• Best for: Production environments needing cloud LLM capabilities
• Key Features:
  • Local infrastructure
  • AWS Bedrock integration
  • Mixed model support (local + cloud)
  • Cost-effective scaling
• Learn More: Hybrid Deployment Guide

3. AWS Mode

• Best for: Enterprise-scale deployments
• Key Features:
  • Full cloud infrastructure
  • Auto-scaling capabilities
  • High availability
  • Enterprise security features
• Learn More: AWS Deployment Guide

4. Kubernetes Mode (Coming Soon)

• Best for: Container-orchestrated environments
• Key Features:
  • Kubernetes native deployment
  • Helm charts
  • Operator support
  • Cloud-agnostic
• Learn More: Kubernetes Deployment Guide

Deployment Comparison

Feature	Local	Hybrid	AWS
Infrastructure	Self-hosted	Self-hosted	AWS Cloud
Model Support	Ollama	Ollama + Bedrock	Ollama + Bedrock
Scaling	Manual	Manual	Auto-scaling
High Availability	No	No	Optional
Security	Basic	Enhanced	Enterprise
Cost	Low	Medium	Usage-based
Setup Complexity	Simple	Moderate	Advanced

Architecture Overview

graph TD
    A[User Request] --> B{Deployment Mode}
    B -->|Local| C[Local Infrastructure]
    B -->|Hybrid| D[Local + AWS Bedrock]
    B -->|AWS| E[Full AWS Infrastructure]
    C --> F[Ollama Models]
    D --> F
    D --> G[AWS Bedrock Models]
    E --> G

Choosing a Deployment Mode

Consider Local Mode if you:

• Need complete data privacy
• Are developing or testing
• Have limited cloud budget
• Want simple setup and maintenance

Consider Hybrid Mode if you:

• Need both local and cloud models
• Want cost-effective scaling
• Have existing infrastructure
• Need flexible deployment options

Consider AWS Mode if you:

• Need enterprise-grade features
• Need security advanced features
• Want automated scaling
• Require high availability
• Have cloud-first strategy

Deployment Requirements

Local Mode

• Docker/Podman
• 8GB RAM minimum
• 20GB storage
• Optional: NVIDIA GPU

Hybrid Mode

• Local Mode requirements
• AWS account
• AWS Bedrock access
• Network connectivity

AWS Mode

• AWS account
• Terraform installed
• AWS CLI configured
• Required AWS services enabled

Getting Started

1. Review System Requirements
2. Choose your deployment mode
3. Follow the corresponding deployment guide:
   • Local Deployment
   • Hybrid Deployment
   • AWS Deployment

Next Steps

• Review Security Considerations
• Set up Monitoring
• Configure Backup Strategy
• Join our Community

---

Horizons OmniChat by evereven