# Nativ

Local AI, native to your Mac. Chat, serve, monitor, and connect MLX models from one macOS app.

Repository: https://github.com/Blaizzy/nativ
Canonical: https://ross.abutalabs.com/products/nativ
Homepage: https://blaizzy.github.io/nativ/
Language: Swift
License: MIT
License Family: permissive
Last push: 2026-08-26T18:17:15+00:00
Link (homepage): https://blaizzy.github.io/nativ/

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 99, longevity 3
- inputs: {"age_days": 44, "days_push": 7, "days_rel": 7, "gap_med": 3.0, "n_releases_24m": 11}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1330, forks 90 (observed 2026-08-28T04:04:22.993523+00:00)

## What it is
Nativ is a free, MIT-licensed macOS desktop application for running open AI models locally on Apple Silicon Macs, built on MLX-VLM. It provides a chat interface with streaming, multimodal support (text, image, video, audio), live performance telemetry, a curated model library, and local endpoints for coding agent integrations.

## Use cases
- run open-source LLMs locally on my Mac without cloud subscriptions
- chat with vision language models and caption images on Apple Silicon
- transcribe audio and summarize video locally
- connect coding agents like Claude Code to a local model endpoint
- monitor tokens per second and memory pressure while running local models
- fine-tune and run VLMs like PaliGemma, Pixtral, and Molmo on a Mac

## When to choose
- you have an M1+ Mac and want private, offline AI inference with no accounts or subscriptions
- you want a polished GUI plus multimodal (text, image, video, audio) local model support
- you need a local OpenAI-style endpoint to power coding agents or other tools

## When to avoid
- you need GPU cluster or CUDA-based inference on Linux servers
- you want the widest possible model ecosystem or vLLM-style production serving
- you are on Windows or Linux, since Nativ is macOS/Apple Silicon only

## Facets
- artifact type: application
- maturity: active
- function: llm-inference, chatbot, speech-recognition, tts, image-processing, video-processing, machine-learning
- domain: large-language-models, artificial-intelligence, machine-learning, self-hosted, privacy, desktop-applications
- platform: -
- tags: local-ai, mlx, apple-silicon, vision-language-models, multimodal, on-device-inference, open-source, chat-ui, macos, desktop

## Member repositories
- Blaizzy/nativ (main) score 80
- Blaizzy/mlx-vlm (backend) score 87

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:22.993523+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 2026-08-29T17:56:56.439434+00:00, confidence not recorded.
  - readme: https://github.com/Blaizzy/nativ (fetched 2026-08-28T04:04:22.993523+00:00, sha c57b96572e71)
  - homepage: https://blaizzy.github.io/nativ/ (fetched 2026-08-29T08:52:12.894652+00:00, sha 6dc0aa7bcb31)
  - registry_pypi: https://pypi.org/pypi/mlx-vlm/json (fetched 2026-08-29T08:52:12.897407+00:00, sha c15320b29828)
- Data as of 2026-08-30T08:39:29.467469+00:00.