Technology & System Requirements
Technical specifications and security architecture of LokalKI.
Data Flow – 100% Local
Your Data
→
LokalKI Server
↔
Employee
✕
No Cloud Data Transfer
Technical Specifications
| Requirement | Details |
|---|---|
| Docker | Container-based deployment, Docker Compose |
| Linux Support | Ubuntu 20.04+, Debian 11+, RHEL 8+ |
| NVIDIA GPU | Recommended for LLM inference (CUDA 11.8+) |
System Requirements
Docker Support
LokalKI is deployed as Docker containers. Full orchestration via Docker or Docker Compose. Easy deployment without complex dependencies.
Hardware Recommendations
- NVIDIA GPU recommended (for optimal inference performance)
- 16 GB RAM minimum (32 GB recommended for larger models)
- Sufficient disk space for vector database and document index
Security Architecture
Air-gapped Capable
LokalKI can run in a fully air-gapped environment. No dependency on external services – ideal for highly sensitive environments.
No External API Calls
All AI inference runs locally. No data is sent to OpenAI, Anthropic or other cloud providers. Full GDPR compliance.
Local Vector Database
Document embeddings and RAG indexes are stored exclusively on your servers. No cloud storage, no third-party infrastructure.