FractalCore schedules training tasks across mobile nodes, averages their checkpoints with FedAvg, and pays them in "Liquid MBs" — plus it slices Llama-class models into phone-sized pieces.
AnyJev turns any open LLM into a calibrated decision engine—typed choices with probabilities you can threshold—without touching a single model weight.
It packages a C++17 video analytics runtime, a browser-based pipeline editor, and async VLM nodes into a single deployable appliance stack for edge hardware.
ollaya applies Ollama's pull-and-serve ergonomics to decision models — small networks that return calibrated probabilities in a single forward pass and never generate a token.
Project Maya runs GLM-5.3-Flash privately on one to sixteen GPUs by keeping hot experts in VRAM, warm ones in RAM, and the rest on NVMe.
TensorFold serves a short list of open model families on Apple Silicon and NVIDIA GPUs behind an OpenAI-compatible API, and accepts a drafted token only when it equals what the same engine would emit serially.
It rewrites every kernel for AMD Strix Halo and a single 179B-parameter MoE family to beat general-purpose engines without spending fewer bits.
opencode2api fronts OpenCode Zen and Zen Go with Chat Completions, Responses, and Anthropic Messages endpoints, plus key pools, proxy rotation, and a built-in WebUI.
comfy-cli treats ComfyUI like the server software it is: installable, scriptable, and manageable from a terminal instead of a browser tab.
To prove that a 2.78-trillion-parameter model can run on a single CPU with 8 GB of RAM and no GPU.
Gufo ditches portability on purpose: a C++ inference engine built solely for AMD's Strix Halo APU, where 128 GiB of unified memory makes a dedicated engine worth writing.
A local-first AI agent that persists its identity across restarts and treats rewriting its own source code as a first-class feature.
A tutorial that bridges the gap between inference theory and production code by making you build a working mini-sglang first.
Because shipping voice features on Mac and iOS shouldn’t require cloud contracts or API keys.
Go CLI that replaces guesswork with an interactive picker for model quantizations, branches, and diffusers components.
Magnitude is an open-source agent that bundles its own local-model runner so you can skip the Ollama setup and run fully offline.
It brings Laya's typed-decision models to MLX so Apple Silicon can score, classify, and route structured inputs locally without generating text.
Dream Server exists because most people would rather pay OpenAI than spend a weekend hand-wiring Docker configs for local LLMs, RAG, and image generation.
To house the code, checklists, and profiling recipes for an O’Reilly guide that treats AI performance as a full-stack systems problem rather than a benchmarking contest.
hls4ml compiles neural networks into FPGA firmware for sub-microsecond inference.
heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.
