# MLC LLM

> Apache-2.0 compiler and deployment engine for running LLMs natively on NVIDIA, AMD, Apple, Intel, web, iOS and Android.

- id: `mlc-llm`
- category: Tools (`tools`)
- url: https://llm.mlc.ai/
- repo: https://github.com/mlc-ai/mlc-llm
- docs: https://llm.mlc.ai/docs/
- status: active
- pricing: open-source
- license: Apache-2.0
- tags: inference, compiler, local-models, mobile, open-source
- auth: none
- access notes: Open-source project. Documented install is nightly wheels via python -m pip install --pre -U -f https://mlc.ai/wheels … (not on PyPI as mlc-llm). No “latest” GitHub Release on 2026-11-14. Commands were not run.
- related: MLX LM (https://indexagentica.com/entries/mlx-lm/); vLLM (https://indexagentica.com/entries/vllm/); SGLang (https://indexagentica.com/entries/sglang/); llama.cpp (https://indexagentica.com/entries/llama-cpp/); LocalAI (https://indexagentica.com/entries/localai/)
- sources: https://llm.mlc.ai/ , https://llm.mlc.ai/docs/ , https://llm.mlc.ai/docs/install/mlc_llm , https://mlc.ai/wheels/ , https://github.com/mlc-ai/mlc-llm , https://github.com/mlc-ai/mlc-llm/blob/main/LICENSE , https://github.com/mlc-ai/mlc-llm/blob/main/README.md
- added: 2026-11-14
- updated: 2026-11-14
- last verified: 2026-11-14
- maintainer: MLC
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/mlc-llm/
- json: https://indexagentica.com/api/entries/mlc-llm.json

Established open project (repo mlc-ai/mlc-llm; not archived; created 2023-04-29; last pushed 2026-11-13; not presented as a 2026 launch). Root LICENSE is Apache-2.0. Site https://llm.mlc.ai/ (HTTP 200). Docs https://llm.mlc.ai/docs/ were HTTP 200; https://llm.mlc.ai/docs 301s there and is not used. The install page title still reads “mlc-llm 0.1.0 documentation”. The GitHub Releases API had no “latest” release (404) on 2026-11-14. There is no PyPI project named mlc-llm (404). PyPI mlc-ai 0.1 (“Machine learning for systems”, homepage github.com/mlc-ai) is a different package and is not the install path for this project. Install docs (https://llm.mlc.ai/docs/install/mlc_llm, HTTP 200) describe nightly prebuilt pip wheels from https://mlc.ai/wheels/ (200; the URL without a trailing slash 301s and is not used), for example: python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-cpu mlc-ai-nightly-cpu, plus platform-specific wheels (cu128, cu130, rocm61, rocm62 and others on that page). A conda environment is recommended. Build-from-source instructions are also documented. None of those commands were executed. Quick start shows MLCEngine and `mlc_llm serve` with Hugging Face MLC model IDs. Model weights and any GPU or host resources are licensed and billed separately.
