GoKawiilRedis creator releases ds4, a local inference engine for frontier open-weight LLMs
Salvatore Sanfilippo, the creator of Redis, has released DwarfStar 4 (ds4), a lightweight C-based inference engine for running large open-weight language models locally on high-memory Mac, CUDA and ROCm hardware. The tool supports models including DeepSeek V4, V4.1 Flash, GLM 5.x and Qwen3.8 Flash Next, covering text and vision, with a local API, CLI and native agent, and is released under the MIT license.
GoKawiil's interpretation of the reporting above, not reported fact.
The project could make it easier for developers to run large, capable models entirely on their own hardware rather than relying on cloud APIs, which matters for privacy, cost and offline use cases. Backing from a well-known systems programmer may draw attention and contributions to the open local-inference ecosystem, though its practical performance against established tools like llama.cpp remains to be seen.
- ds4 is a new local inference engine from Redis creator Salvatore Sanfilippo
- Supports frontier open-weight models: DeepSeek V4/V4.1, GLM 5.x, Qwen3.8 Flash Next
- Built in C with Metal, CUDA and ROCm support, MIT licensed, runs Qwen on 64GB machines
Apple Mac Studio (M2 Ultra) — Running frontier open-weight models like DeepSeek V4 or GLM 5.x locally demands serious unified memory, and the Mac Studio's high-RAM configurations are built exactly for that kind of workload. Its Metal-accelerated architecture pairs well with ds4's native Metal support, letting you load large quantized models without needing a discrete GPU rig. It's a practical way to get frontier-scale local inference on your desk.
See Apple Mac Studio (M2 Ultra) on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: dwarfstar.sh — Dwarfstar Community, 2026-10-02
Published there as: “From the creator of Redis; run LLM locally with ds4”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.