Real home AI setups from RunAIatHome users. See which GPUs people are actually using to run local models, what RAM and storage they chose, and how it's working for them.
Share your local AI setup with the community. Submissions are reviewed before publishing. Include your GPU, RAM, which models you run, and any tips from your experience.
Submit via GitHub DiscussionsModels running
Running Ollama + ComfyUI side-by-side. 70B model fits fully in VRAM with room to spare. Whisper transcribes meeting recordings in real-time.
Affiliate disclosure: Links below are sponsored.
Models running
Q4 quantization lets me run 8B models at full speed. SDXL takes ~12s per image. Great value build for under 700€ total.
Affiliate disclosure: Links below are sponsored.
Models running
The 16GB VRAM is the sweet spot for 13B models. CodeLlama answers complex refactoring questions. Context window is the real limit, not VRAM.
Affiliate disclosure: Links below are sponsored.
Models running
Budget-friendly entry point. Phi-3 Mini is surprisingly capable for daily tasks. I transcribe all my voice memos with Whisper — privacy matters to me.
Affiliate disclosure: Links below are sponsored.
Use our free VRAM Calculator to find out which models run on your current GPU, or the GPU Comparator to compare cards side by side.
Open VRAM CalculatorFound this useful? Get guides like this in your inbox every week.
