# mlc-ai/mlc-llm

Universal LLM Deployment Engine with ML Compilation

Repository: https://github.com/mlc-ai/mlc-llm
Canonical: https://ross.abutalabs.com/products/mlc-llm
Homepage: https://llm.mlc.ai/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: llm, machine-learning-compilation, language-model, tvm
Last push: 2026-08-17T20:50:06+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 8, longevity 87
- inputs: {"age_days": 1223, "days_push": 16, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 23095, forks 2124 (observed 2026-08-28T04:11:34.426280+00:00)

## What it is
MLC LLM is a machine learning compiler and high-performance deployment engine for large language models, built on TVM-based ML compilation. It provides a unified OpenAI-compatible inference engine (MLCEngine) that runs LLMs natively across GPUs, CPUs, browsers, and mobile platforms.

## Use cases
- run llama 3 locally on my gpu
- deploy an llm on android or ios
- serve an openai-compatible api for a local model
- run a chat model in the browser with webgpu
- quantize and compile an llm for apple silicon
- host a local llm rest server
- run llm inference on amd or intel gpus

## When to choose
- you need to deploy LLMs across diverse hardware (NVIDIA, AMD, Intel, Apple GPUs, mobile, browser)
- you want an OpenAI-compatible API backed by a local engine
- you need int4-quantized models running on consumer devices
- you want a single engine for Python, JavaScript, iOS, Android, and REST serving

## When to avoid
- you only need a simple CPU-only inference server with minimal setup
- you want the broadest ecosystem of prebuilt community models rather than MLC-compiled weights
- you need training or fine-tuning rather than inference deployment

## Facets
- artifact type: framework
- maturity: active
- function: llm-inference, machine-learning, compiler, http-server, sdk
- domain: large-language-models, machine-learning, gpu-computing, cross-platform, developer-tools
- platform: windows, wasm, python, cli
- tags: ml-compilation, tvm, openai-compatible-api, vulkan, metal, webgpu, local-llm, quantization, linux, macos, android, ios, web, gpu

## Member repositories
- mlc-ai/mlc-llm (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:34.426280+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 2026-08-29T16:56:47.845246+00:00, confidence not recorded.
  - readme: https://github.com/mlc-ai/mlc-llm (fetched 2026-08-28T04:11:34.426280+00:00, sha dfdc0cfaa898)
  - homepage: https://llm.mlc.ai/ (fetched 2026-08-29T07:54:42.887742+00:00, sha c2ad3ccc7ec8)
  - site_page: https://llm.mlc.ai/docs (fetched 2026-08-29T07:54:42.896942+00:00, sha ae85a95b0c49)
  - site_page: https://llm.mlc.ai/docs/get_started/quick_start (fetched 2026-08-29T07:54:42.898960+00:00, sha 58c1b90f013a)
- Data as of 2026-08-30T08:39:29.467469+00:00.