ModelDeck Turns Closet GPUs Into a Phone-Controlled Local AI Hub, Streaming Multiple Models via Tailscale
EAccelerate_42 · x · 2026-09-05
A developer demoed ModelDeck, a phone app that controls a home GPU server in his closet: load, swap and stream multiple local models with live VRAM/temperature monitoring. It supports DeepSeek V4 Flash (384K context), Qwen3.8 27B (NVFP4), GLM-5.3 Flash, plus Qwen-Image editing and MiniMax H3 video with sound. Access is Tailscale-only — no cloud, no token bill; early access preview coming soon.
More from Infra
- Jensen Huang: 1GW of AI data center costs $50-60B, and 'we're building 100 gigawatts' by decade's end — victor_explore · 2026-09-05
- Full recipe: running Qwen3.8 27B on AMD Strix Halo with patched ROCm llama.cpp — ilintar · 2026-09-05
- Open-sourced Lightpanda: headless browser 11x faster than Chrome with 9x less RAM — JafarNajafov · 2026-09-05
- Tencent Hunyuan preview: 770B params, 1M context, Apache 2.0, 214GiB quantized — Aiden_Tech_Ai · 2026-09-05
- NVIDIA's $1B Single-Model Training Forecast Landed Years Ahead of Schedule — IgorCarron · 2026-09-05
- Starting local AI on attic hardware: 3x NUC11 plus Ryzen 3700X rig — -markusb- · 2026-09-05