The Personal Computer for AI Agents

Lucebox builds plug-and-play hardware and the open inference stack that powers it. Local-first, OpenAI/Anthropic-compatible, 4 to 6x faster than competing boxes at the same price.

Lucebox is a 9.56L aluminum box that runs frontier open models locally at speeds people thought needed cloud GPUs. Inside: an RTX 3090 (24GB GDDR6X) paired with an AMD Ryzen AI MAX+ 395 APU (128GB unified LPDDR5X), 2TB NVMe, Corsair 750W 80+ Gold. Outside: a single power cable and an OpenAI/Anthropic-compatible endpoint reachable from Claude Code, Codex, OpenCode, Hermes, OpenClaw, Open WebUI, and Ollama in roughly one minute from unboxing.

The speed comes from this org. lucebox-hub is the open inference engine: custom CUDA kernels, speculative decoding (DFlash), speculative prefill compression (PFlash), and persistent megakernels. It runs on any RTX 30/40/50, on Strix Halo, and on Radeon 7900 XTX, not just on our hardware.

• Privacy. Prompts and weights never leave the device. Default-fit for legal, medical, finance, and any team where the data is the moat.
• Cost. $4,900 once, then zero per token. Replaces $200 to $2,000 per month in cloud API spend for sustained agent workloads.
• Throughput. Up to 207 tok/s on Qwen3.5-27B and 134 tok/s at 128K context, matching or beating cloud latency on a desk.
• Open. Apache 2.0 inference stack, GGUF models, no vendor lock-in.

Repository	Description	Stars	Forks
lucebox-hub	Fast LLM speculative inference server for consumer hardware. DFlash + PFlash + Megakernel. OpenAI/Anthropic compatible HTTP server.
llama.cpp-dflash-ggml	llama.cpp fork with DFlash speculative decode and ggml DDTree integration. Upstream-tracking.

git clone --recurse-submodules https://github.com/Luce-Org/lucebox-hub
cd lucebox-hub
cmake -B server/build -S server -DCMAKE_CUDA_ARCHITECTURES=86
cmake --build server/build --target dflash_server -j

./server/build/dflash_server \
  model.gguf \
  --draft draft.gguf \
  --port 8000

Then point any OpenAI-compatible client at http://localhost:8000/v1.

• NVIDIA: RTX 3090, RTX 4090, RTX 5090, RTX 2080 Ti (CUDA 12+)
• AMD: Ryzen AI MAX+ 395 Strix Halo (HIP / ROCm 6+), RX 7900 XTX (HIP)
• OS (inference engine): Linux, Windows
• OS (Lucebox appliance): Linux, pre-tuned

The plug-and-play Lucebox ships pre-tuned with the full stack loaded. $4,900, one year warranty, refurbished and fully serviced RTX 3090.

→ lucebox.com

• Product Site
• Blog
• DFlash on 27B
• PFlash Speculative Prefill
• Megakernel Decode
• Laguna-XS.2 @128K
• Gemma vs DeepSeek
• Client Harnesses
• AMD Strix Halo Notes
• eGPU Myth
• Issue Tracker

Website • GitHub • X • Discord

_{Apache 2.0. Built in Italy.}