I will setup self hosted local ai stack on your nas or home server

Z
zoren_ban
Z
zoren_ban
Zoren B

About this gig

Tired of paying $20/month for ChatGPT Plus, or worried about sending private photos to Midjourney?

I'll deploy a complete self-hosted AI stack on your hardware 100% private, $0/month after setup.


What you get (Basic $120):

- Ollama (latest) + Llama 3.1 8B quantized (Q4_K_M, runs on 8GB RAM)

- Open WebUI (ChatGPT-like interface, multi-user)

- Web search integration (optional)

- API access for your apps (compatible with OpenAI SDK)


Add-ons (Standard $250):

- ComfyUI + Stable Diffusion XL (image generation, runs on 6GB+ VRAM)

- RAG pipeline: ChromaDB + LangChain + PDF/URL ingestion

- 5 sample LoRA adapters


Premium ($400):

- 24/7 monitoring + auto-restart script

- Daily backup of models + vector DB to S3/NFS

- Custom LoRA training on your data (1 epoch, 1 model)

- 30-day support

Hardware check:

- 8GB RAM minimum (CPU-only Llama 3 8B)

- NVIDIA 6GB+ VRAM for SD (RTX 3060+ recommended)

- 50GB+ disk for models

Why self-host:

- 100% private (no data sent to OpenAI/Stability)

- Save $240/year (ChatGPT Plus + Midjourney Basic)

- No censorship / no API rate limits

- Customize with your data (RAG)

Requirements:

1. SSH access to your server

2. 8GB+ RAM (16GB recommended)

3. (Optional) NVI

Get to know Zoren B

Zoren B

NAS Engineer,Docker Gitea Tailscale Uptime Kuma Setup

  • FromChina
  • Member sinceOct 2026
  • Avg. response time3 hours
  • Languages

    English, Chinese
I am a self-hosted lab engineer and consultant based in Shanghai, specializing in NAS infrastructure and Docker orchestration. I manage over 30 containers daily, including Gitea, Tailscale, and AI stacks like Ollama and Open WebUI. I help users transition from paid SaaS to private, self-hosted alternatives.