Karyam On-Prem
Karyam can be deployed entirely within your own infrastructure.
Supported platforms:
- Linux
- macOS
- Windows
Deployment packages include:
- Application binaries
- Upgrade utilities
- Configuration files
System Requirements
Section titled “System Requirements”Before installing Karyam, ensure your environment meets the following minimum requirements.
| Component | Minimum Requirement | Recommended |
|---|---|---|
| CPU | 4 Cores | 8+ Cores |
| Memory | 8 GB RAM | 16 GB+ RAM |
| Operating System | Linux, macOS, or Windows | Latest stable version |
| Architecture | AMD64 (x86_64) | AMD64 (x86_64) |
| Network | Internet access for downloads | Stable broadband connection |
Optional Components
Section titled “Optional Components”Depending on your deployment, you may also configure:
- PostgreSQL
- Vector Database (PostgreSQL pgvector or Qdrant)
- External AI Model Providers (OpenAI, Anthropic, Gemini, etc.)
- Ollama for local models
Supported Platforms
Section titled “Supported Platforms”Karyam supports deployment on Linux, macOS, and Windows using the same installation workflow.
Select your operating system and Karyam version below.
Downloads
Choose the package that matches your operating system and CPU architecture.
Extract the package
Enter the installation directory
Configure Karyam
Edit karyam.toml and configure the host and port for your environment.
host = "0.0.0.0"
port = 8080Start Karyam
Deploy Karyam on Linux
Follow the complete installation process step by step.
Upgrade an existing installation
Existing installations can be upgraded without affecting workspaces, models, flows or vector database configuration.
./data/bin/upgrade.sh.\data\bin\upgrade.ps1The automated upgrade routine securely preserves the following directories and dynamic local data:
- Workspaces and asset storage
- Custom LLM Agents config profiles
- Automated workflows & logic flows
- Cached local weights and models
- System and Vector Database configurations
Deployment Models
Section titled “Deployment Models”Karyam can be deployed in:
- Private Datacenters
- Enterprise Infrastructure
- Virtual Machines
- Bare Metal Servers
- Cloud Providers
The installation process remains identical regardless of where Karyam is deployed.
