SkillFishOS can run local LLMs on the BC-250's integrated GPU, accelerated in Vulkan — nothing leaves the machine.

• Unsloth Studio as the engine, accelerated via Vulkan on the Cyan Skillfish GPU. It is a single native service that provides both the chat interface and an OpenAI-compatible API, listening on 127.0.0.1:8888.
• A one-click AI panel (skillfish-ai-panel) that starts and stops it.
• The Remote Manager proxies it at /unsloth, so the same engine is reachable from another machine on the LAN without opening a second port.

It used to be Ollama plus OpenWebUI in three Docker containers with a ~6.5 GB custom image. That is gone since 26.08: one service, no Docker.

Why Vulkan matters here: measured on this hardware, the same model runs at about 210 tokens/s on the GPU against 41 on the CPU. It is not a rounding difference, it is the whole reason on-device AI is usable on this board.

• A first-run wizard installs the stack and lets you pick a model from a curated list of 30+ options that fit the hardware (≤14B).
• Live readout of CPU / GPU / VRAM / RAM.
• A slider to grow the shared memory (GTT) available to the model — see Memory VRAM and GTT.
• Turn the engine off with one click to free the GPU and memory for gaming.

• Bigger models need more memory: raise the GTT budget (and/or the UMA VRAM) before loading a large model.
• The GPU governor's idle behaviour means the card drops to 350 MHz between prompts; under inference it ramps up. For sustained throughput you can pick Performance GPU Governor and Tuning — though for pure compute the 2230 MHz point can help (with adequate voltage/cooling), the shipped safe cap is 2200 MHz.
• Unlocking Compute Units (40-CU) roughly ×ばつ's the GPU's compute throughput.

Nothing here phones home — the models and the chat run entirely on your BC-250.

Two repositories by akandr are worth reading if you want to go further on this board:

• bc250 — setting the board up for Ollama and Vulkan inference, plus image generation with stable-diffusion.cpp.
• bc250-rocm — ROCm/HIP made to work on gfx1013, with the driver patches, the recipe and the measurements.

The second one answers a question we are asked often: why Vulkan and not ROCm. Its comparison puts ROCm decode at roughly 40 to 63 percent of Vulkan across most models tested, with one model close to parity. That measurement was taken on someone else's board, which is exactly what makes it worth citing here.

It also carries a warning worth repeating. Under sustained load a GPU page fault can escalate to a GPU reset, and on this hardware the reset takes the machine with it. The captured trace shows the reset itself succeeding, and the host then stalling on a clocksource watchdog. This is not a board to leave running unattended on work that matters.