Skip to main content
Local LLM runs language models on your Mac’s Apple Silicon, with no account or API key. After the model downloads, chats with it stay on your Mac. It needs macOS 15 or later.

​
Set up

1

Open the provider

Open Settings → AI Model → Local LLM (MLX).
2

Pick a model

Open the Model list. Enconvo suggests a model that suits your Mac, and you can search Hugging Face for others.
3

Download it

Start the download and wait for it to finish. Enconvo checks free disk space first, and a download resumes after a restart or a network drop.
4

Test it

Send a short message. The first reply takes longer while the model loads into memory.

​
Models and options

To see how much memory and disk your local models use, or to delete one, open Settings → Local Models. Models load again when a feature needs them.

​
Troubleshooting

Local models

Keep a whole task on your Mac, not only the chat model.

Ollama

Use models served by Ollama.

LM Studio

Use models served by LM Studio.

Provider catalog

Local providers for speech, OCR, and embeddings too.