I'm new: start here

Get a model running on your computer, then ask it to draft an email.

A model is a file. A chat app opens that file and lets you talk to it. This guide uses LM Studio, a separate app.

1. 1. Find a model that fits

   This walkthrough uses Qwen2.5-1.5B-Instruct-GGUF, a small chat model. Check the app requirements above first. This model has not been tested on your computer.

   Want more choices? Use the finder, choose “Ask questions and write”, then select your computer's memory.

   Find a model for my computer

   A file-size match is a starting point. It does not prove a model will run well or write good emails. For a file you already chose, use its file-opening help below.

2. 2. Get the app and model

   1. Download LM Studio for Linux, install it and open it.
   2. Open Discover. Search for Qwen2.5-1.5B-Instruct-GGUF, or paste the model's Hugging Face address.
   3. Open the file choices. Choose Q4_K_M only if the app marks it as supported by your computer. This is a smaller version of the model.
   4. Click Download and wait until it finishes. The model downloads from its publisher, rather than Hugging Bay.
   Already downloaded a model file from Hugging Bay?

   Keep the exact filename. The file-opening guide explains how to import it. Double-clicking the file may not open a chat.

   Open a downloaded model file

   After importing, continue to step 3 below.

3. 3. Run it on your computer

   1. Open Chat in LM Studio.
   2. Use the model picker at the top to select the model you downloaded. Wait for it to finish loading.
   3. Type the example below into the message box and send it.

   Write a short, friendly email asking to move our meeting to Friday.

   You're running a model when its reply appears in the chat. Read the draft and check its details before using it.

   If it does not load, close other apps and try a smaller file. If the app reports unsupported hardware, check its requirements or use another supported computer.

   Help if the model will not load