Getting Started¶
Get Sekha up and running in minutes.
Quick Navigation¶
1. Quickstart¶
Fastest way to try Sekha (5 minutes):
- Docker Compose one-liner
- Test with sample conversation
- Verify installation
2. Installation¶
Complete installation options:
- Docker Compose (recommended)
- Pre-built binaries
- Build from source
- Cloud deployment (coming soon)
3. Configuration¶
Configure Sekha for your needs:
- Environment variables
- Config file options
- LLM provider setup
- Storage configuration
4. First Conversation¶
Store and retrieve your first memory:
- Create a conversation
- Perform semantic search
- Assemble context for LLM
What You Need¶
Minimum Requirements¶
- OS: Linux, macOS, or Windows with WSL2
- Docker: 20.10+ with Docker Compose
- RAM: 4GB minimum, 8GB recommended
- Storage: 10GB available
- CPU: 2 cores minimum
Recommended Setup¶
- RAM: 16GB for Ollama models
- Storage: 50GB+ for models and data
- CPU: 4+ cores for better performance
- GPU: Optional, for faster embeddings
Deployment Options¶
Development (Local)¶
Best for: - Learning Sekha - Development - Personal use
Production (Server)¶
Best for: - Team deployments - Public APIs - High availability
See Production Guide for details.
Cloud (Coming Soon)¶
Managed deployment options:
- AWS - Coming Q1 2026
- Azure - Coming Q1 2026
- GCP - Coming Q1 2026
- DigitalOcean - Coming Q2 2026
Learning Path¶
For Developers¶
- Quickstart - Get running
- First Conversation - Basic usage
- API Reference - Learn the API
- Python SDK - Use the SDK
- AI Coding Assistant - Build something
For Users¶
- Quickstart - Installation
- Claude Desktop - MCP setup
- First Conversation - Try it out
- FAQ - Common questions
For DevOps¶
- Installation - Deployment options
- Configuration - Settings
- Production Guide - Production setup
- Security - Hardening
Next Steps¶
Start with the Quickstart Guide →
Or explore:
- Architecture - How it works
- Integrations - Connect to tools
- Guides - Use case examples