MiniPC Server
Self-hosted server dashboard with local AI capabilities
The Problem
Managing a personal server requires SSH access and terminal knowledge. Running AI models locally needs complex setup and monitoring.
The Solution
A web dashboard that simplifies server administration and provides a visual interface for running local LLMs — making self-hosting accessible without terminal expertise.
Architecture
Next.js web interface for server management and AI chat
Node.js backend running on the MiniPC — handles system commands and proxies AI requests
Ollama for running LLM models locally — privacy-first, zero API costs
System metrics collection: CPU, RAM, disk, network, and service health
Key Features
System monitoring
Real-time CPU, RAM, disk, and network metrics with historical charts
Docker container management
Start, stop, restart, and inspect Docker containers from the web UI
Local LLM chat interface
Chat with Ollama-powered local models — complete privacy, no data leaves the server
Service health dashboard
Monitor uptime and health of all running services with alert notifications
Remote terminal
WebSocket-based terminal for advanced administration without SSH client
+2 more features in the repo
Under the Hood
Local AI Pipeline
User prompts go to Ollama API running locally, responses stream back through WebSocket to the browser — no cloud API, no latency from network round-trips, full data privacy.
- User prompt
- Ollama API
- Streaming response
- Web display
Docker Orchestration
The backend communicates with the Docker daemon via the API to list, start, stop, and inspect containers — providing a visual interface over the Docker CLI.