workflowstacks
design

mlxcel

High-performance LLM, VLM, embedding, reranking, and audio inference for Apple Silicon and NVIDIA CUDA systems.
473 stars60 forksRustUpdated 10/22/2026100% free · open source

About this skill

High-performance LLM, VLM, embedding, reranking, and audio inference for Apple Silicon and NVIDIA CUDA systems.
Use with Claude
New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

🛠️ Technical setup

Expect 20–40 minutes in a terminal — or let your AI agent drive it.

Try it instantly — no install
Claude Code
mkdir -p ~/.claude/skills/mlxcel && curl -fsSL https://workflowstacks.com/api/skills/mlxcel/claude-skill -o ~/.claude/skills/mlxcel/SKILL.md
Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)
claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How mlxcel works
Codeflow
Free to inspect

Mlxcel is a very large Rust project (~1.2M lines across 2935 code files, plus 211 test files). It is a full software project: use it through its install path rather than reading it end to end. Last commit this month, Apache-2.0 license, has a test suite.

Size
Very large codebase
~1.2M lines · 2935 code files · days to read — use, don't read
Setup
Developer setup
A real software project. Use it via its install path; don't expect to read it all.
Runs on
Rust
No API keys detected
Rust 78%MLIR 7%C++ 5%TypeScript 3%Python 3%
Where to start reading
  1. 1
    README.md
    Start here — what it does and how to install it
  2. 2
    src/main.rs
    Where the program starts running
  3. 3
    Cargo.toml
    Dependencies and the commands it exposes
  4. 4
    examples/async_bench.rs
    A worked example — copy this to get going
What's in each folder
src/Core code — the actual logic2448 files
docs/Documentation1943 files
examples/Examples you can copy54 files
recipes/Examples you can copy6 files
TECHNICAL_REPORTS/Folder837 files
benchmarks/Evaluations & benchmarks605 files
webui/Folder250 files
scripts/Helper scripts131 files
READMEHas testsDocumentedCI checksExamples includedApache-2.0 licenseUpdated this month
Quick Actions
Details
Creator
lablup
Language
Rust
Category
design
Published
1/30/2026

Are you the creator of this tool? and earn 85% of every sale.

Show it off in your README:

Featured on WorkflowStacks
[![Featured on WorkflowStacks](https://workflowstacks.com/api/badge/mlxcel.svg)](https://workflowstacks.com/skills/mlxcel?utm_source=github&utm_medium=badge)
🔥 Hot this week, in your inbox
Skills like this, every Monday.

The five fastest-growing open-source AI skills, ranked by GitHub star growth. One email a week, unsubscribe anytime.

Related skills

More design tools founders pair with this one.

design★ 392,732
openclaw: AI assistant that runs on any OS and protects your data
Backed by 391k+ GitHub stars, openclaw gives founders a cross‑platform personal AI assistant that runs locally, keeping your data under your control.
design★ 165,093
Stable Diffusion web UI for Design
Generate images with stable-diffusion-webui, a tool for founders. 163k+ GitHub stars.
design★ 113,119
Supabase: Rapid Postgres‑Backed Apps for Founders
Launch web, mobile, or AI products on a fully managed Postgres database with real‑time APIs and auth—ideal for startup teams. 111k+ GitHub stars.
design★ 5,313
rallly
Rallly is an open-source scheduling and collaboration tool designed to make organizing events and meetings easier.
design★ 15,101
ImageToolbox: Fast, Advanced Image Editing for Mobile Apps
Empower your Android app with ImageToolbox’s 15k+ GitHub stars, offering cropping, drawing, filters, OCR, and background removal—perfect for founders building photo‑heavy experiences.
design★ 16,233
labelme: Fast Image Annotation for Vision Projects
Build accurate polygon, rectangle, circle, line or point annotations—and AI‑assisted labeling—in Python. Ideal for founders developing computer‑vision models.