@tom_doerr: Personal AI Computer build guides with up to 384GB VRAM https://github.com/autonomous-ai/autonomous-computer…

X AI KOLs Timeline Tools
personal-ai-computer build-guide open-source nvidia-gpu local-ai hardware vram

Summary

Open-source build guides for a personal AI computer with up to 384GB VRAM, supporting configurations from home to on-prem business. Includes bill of materials, assembly photos, and software setup for running open-source AI models locally.

Personal AI Computer build guides with up to 384GB VRAM https://github.com/autonomous-ai/autonomous-computer…
Original Article
View Cached Full Text

Cached at: 07/02/26, 02:24 PM

Personal AI Computer build guides with up to 384GB VRAM https://github.com/autonomous-ai/autonomous-computer…

autonomous-ai/autonomous-computer

Source: https://github.com/autonomous-ai/autonomous-computer

Autonomous Computer — Build Your Own Personal AI Computer

https://github.com/user-attachments/assets/3e410e5d-83f4-4aed-a8b4-2426781f3ebd

License: MIT GitHub stars GitHub forks Status Repo Size Open Issues
Own your intelligence.
Open-source guides to build a machine that runs open models on hardware no one can switch off — every part, every bracket, every BIOS setting, every assembly photo. Pick the size that fits your budget and your work.

Why this exists. The best models live in someone else’s cloud, behind someone else’s terms and someone else’s government. A model you rent can be cut off overnight. A model running in your own house can’t. Build it once; own it for good.

Pick your configuration

One Personal AI Computer, three configurations — from a home rig to an on-prem business cluster. Each is a complete, self-contained guide: bill of materials, 3D-printable/CNC housing files, wiring, BIOS, and assembly photos.

ConfigGPUsVRAMBest forPlatformGuide
2×2× RTX 509064 GBHome — your first local rig, runs a real model tonightIntel Xeon W5 · ASUS W790→ 2× config
4×4× RTX PRO 6000 Blackwell384 GBTeam — big models + agents running all dayAMD EPYC 9124 · ASRock Rack→ 4× config
8×8× RTX 4090 / 5090192–256 GBOn-prem business — a company’s AI on its own floor; IP and data never leaveDual AMD EPYC 9004 (Genoa)→ 8× config

What it can run

VRAM is the constraint that decides which open models fit. Rough guide (exact fit depends on quantization — model specifics live in /software):

BuildVRAMOpen models it can serve (e.g.)
2× RTX 509064 GBLlama 70B / Qwen 72B (quantized), 30B-class at full precision, coding + agent models
4× RTX PRO 6000384 GBDeepSeek-V3 / R1, Qwen 235B, Llama 405B (quantized) — frontier-class open weights
8× RTX 4090/5090192–256 GBMultiple large models served at once, plus headroom for fine-tuning

Quick start

  1. Pick a configuration above by budget and the models you want to run.
  2. Source the parts from that config’s Bill of Materials.
  3. Make the housing — print the STL files or CNC the STEP files in that config’s folder.
  4. Assemble — follow the config’s photo-by-photo assembly guide.
  5. Set up the software — drivers, BIOS, and serving open models locally: /software.
  6. Run your intelligence — point your agent at localhost and never get cut off again.

The configurations

Software

Bringing the box to life — OS, NVIDIA drivers, BIOS tuning, serving open models (Ollama / vLLM / llama.cpp), and connecting an agent: /software. (Setup and testing are documented today; local model-serving guides are in progress.)

Contributing

Built one? Improved a part? Found a better component? See CONTRIBUTING.md — and share your build. The best community builds get featured.

License

Open source under the MIT License. Fork it, change it, build your own and sell it — we just want it built.

Autonomous — the AI hardware company.
Questions? Open an issue.

Similar Articles