LM Studio

0. Check if your computer is capable of running a local LLM

You should first visit this website to check what kind of LLM model your computer can run:

LLM Hardware Requirements

Here is an example result from the website:

Generally, you can easily run models with a quality score above 80 on your computer.

If your result does not show any score over 80, you might have difficulty running an LLM locally.

1. Download LM Studio and an LLM Model

Open the website:

LM Studio

Download the version suitable for your computer.

After installation, go to the left panel and click on the Search icon:

In the search panel, you can type in keywords based on the test results from the hardware requirements website:

Click on a model in the search results list, and you can download the LLM to your computer:

Wait for the LLM download process to finish before moving on to the next step.

2. Local Service Setup

Step 1: Load Your Downloaded Model

Before starting the service, you must tell LM Studio which model you want to run.

1. Click the Developer icon on the left sidebar.
2. Click Load Model.
3. Choose the model you just downloaded. Keep all settings at their defaults and click Load Model. Wait a moment for it to load into your computer's memory.

---

Step 2: Start the Local Server

1. In the main window, look for the green button that says "Start Server".
2. Click Start Server. You will see a log window pop up at the bottom showing that the server is now running.
3. Enable local network access.

---

Step 3: Copy and Save Your Essential Credentials

Open a notepad or text document and copy/paste these essential pieces of information from the LM Studio screen:

1. The Base URL (where the server lives)

• What it looks like: http://localhost:1234/v1
• Where to find it: Look at the top right of the local server panel. Click the button in the red frame to copy it.

1. The API Key (the password)

• Why: By default, LM Studio running locally on your computer does not strictly require a secret API key/password. However, if you expose this service to your local network (so anyone on the network can access it), you should set up a password for this service.

---

> 💡 Keep it Running: Keep LM Studio open in the background while using your other apps. If you close LM Studio, the local service will turn off and your other apps won't be able to talk to the AI.

---

Once you have completed these steps, you will have:

• A running LLM service
• The service Base URL
• The service password (authentication token)

3. Configure LibGrow

On the LibGrow App:

1. Open the AI provider settings.
2. In the AI provider section, select the LM Studio configuration.
3. Set the LM Studio service Base URL.
4. Set the password (authentication token).

Example:

Base URL: http://192.168.1.25:1234/v1
API Key:  sk-dsadsadsa

Then scroll down to refresh available models:

Finally, select the model from LM Studio:

Troubleshooting

Open any book, and in the AI chat side panel, type a message (e.g., "hi") to verify that the LLM responds without any issues.