# danveloper/flash-moe

Running a big model on a small laptop

Repository: https://github.com/danveloper/flash-moe
Canonical: https://ross.abutalabs.com/products/flash-moe
Language: Objective-C
License Family: other
Last push: 2026-03-19T17:21:57+00:00

## Health v2 (maintenance only)
Score: 48/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 73, release rhythm 35, longevity 12
- inputs: {"age_days": 168, "days_push": 167, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4086, forks 510 (observed 2026-08-28T04:08:34.493522+00:00)

## What it is
An application for running large language models locally on small laptops, written in Objective-C. It enables on-device inference of big models on consumer hardware.

## Use cases
- run a large language model on my laptop
- run big LLM locally without a GPU server
- on-device inference of mixture-of-experts models
- chat with a large model offline on a Mac
- run MoE models on Apple Silicon

## When to choose
- you want to run large models on consumer laptop hardware
- you prefer native macOS performance without cloud dependency
- you need offline, private LLM inference

## When to avoid
- you need multi-GPU or datacenter-scale serving
- you want a cross-platform Linux/Windows-first tool
- you need fine-tuning or training capabilities

## Facets
- artifact type: application
- maturity: active
- function: llm-inference, machine-learning
- domain: large-language-models, machine-learning, gpu-computing
- platform: cpp
- tags: mixture-of-experts, on-device-inference, apple-silicon, quantization, macos, desktop

## Member repositories
- danveloper/flash-moe (main) score 48

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:34.493522+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:23:23.510058+00:00, confidence not recorded.
  - readme: https://github.com/danveloper/flash-moe (fetched 2026-08-28T04:08:34.493522+00:00, sha 6ebdb617a810)
- Data as of 2026-08-30T08:39:29.467469+00:00.