On-device AI execution
Runs requests on your Mac so content does not need to leave the device for local workflows.
Local é um app para macOS que executa AI no seu próprio computador, ajustada ao chip e à memória do Mac para chats, programação, notas de reuniões e fluxos privados no dispositivo.
Local is a macOS app for running AI on your own machine. It is built around the idea that chat, coding, meeting notes, and other common AI tasks can run on-device instead of in the cloud, so your work and memory stay on your Mac by default.
The app uses BaseRT, Base Compute’s inference engine, to tune itself to your chip on first launch. It reads your Mac’s hardware and memory, then recommends models that fit and runs them locally. The product also supports optional cloud models through your own API key, but those requests are separate from Local’s privacy mode.
Runs requests on your Mac so content does not need to leave the device for local workflows.
Reads your chip and memory to suggest models that fit your machine.
Ships with BaseRT, which is compiled for your silicon and designed to improve tokens per second on supported Macs.
Covers file-based chat, in-place coding assistance, and on-device meeting transcription with speaker labels.
Keeps lasting facts about you on the device, with the option to mark sensitive memory so it never leaves the Mac.
Lets you connect GPT/Claude-style providers using your own API key when you need a frontier model.
Ask questions, draft writing, and summarize files or PDFs while keeping the work local to your Mac.
Point Local at a codebase so it can read, edit, and run code in place without uploading the repository.
Transcribe meetings locally and produce written notes with speaker labels.
Store a short list of lasting facts that you can review and edit, including sensitive items that stay on the machine.
Use local models for everyday work and switch to your own-key cloud models only for tasks that need more capability.
Yes. You need an internet connection to download the app and a model, but after that the product can run offline on your Mac.
The homepage says Local supports any Mac with Apple silicon.
No data collection is claimed on the homepage. It says everything lives in a database on your Mac, with no account and no telemetry.
Yes. The homepage says you can connect those models with your own API key, and those requests go to the provider.
The site attributes its speed to BaseRT, its inference engine, which is compiled for your chip rather than shipped as a generic build.
Os dados de tráfego são apenas para referência.
fx.sh
fx é um agente de programação nativo, pequeno e aberto, com CLI para trabalhar com código pelo terminal. Foi desenvolvido para pesquisa, incorporação e uso como parte de sistemas maiores, com suporte à inferência local ou na nuvem.
forte-ai.com
Automação de fluxos de áudio no macOS para Pro Tools e Logic Pro
crewtower.app
Crew Tower é um utilitário para macOS que permite aprovar e acompanhar agentes de programação com AI pelo recorte da barra de menus. Consulte permissões, respostas, progresso e contexto sem alternar para o terminal.
shogunaios.com
ShogunAI é um assistente de memória com IA, privado e focado no processamento local para macOS. Registra o contexto de trabalho no seu Mac para ajudar você a recuperar informações e preparar ações nas suas ferramentas.
charm.sh
Apps CLI e bibliotecas de código aberto para fluxos de trabalho nativos do terminal
deta.surf
Espaço de trabalho local para navegar, fazer anotações e pesquisar com IA