travisvn_obsidian-vision-recall/docs/ollama-setup.md

Setting up Ollama for Obsidian Vision Recall Plugin

This guide will walk you through setting up Ollama on your computer and configuring it to work with the Vision Recall plugin.

Installing Ollama

1. Visit the Ollama website and download the installer for your operating system:

   • macOS (Apple Silicon or Intel)
   • Linux
   • Windows

2. Follow the installation instructions for your system:

   • macOS: Open the downloaded file and drag Ollama to your Applications folder
   • Linux: Run the install script
   • Windows: Run the installer

Starting Ollama

1. macOS & Linux:
   • Open Terminal
   • Run ollama serve
2. Windows:
   • Ollama should start automatically as a service
   • You can verify it's running by opening Command Prompt and typing ollama serve

Pulling a Vision Model

1. Find the model you want to run (and that your system can handle) from the Ollama search
2. Open a new terminal window (while keeping Ollama running in the first one)
3. Run the following command, for example:

   ollama pull llama3.2-vision:11b
   

   This will download the Llama 3.2 Vision model, which is capable of processing both text and images.

Note: The initial download may take several minutes depending on your internet connection.

Configuring the Vision Recall Plugin

1. Open Obsidian and go to Settings

2. Click on "Vision Recall" in the left sidebar

3. Under "LLM settings":

   • Set "LLM provider" to "Ollama"
   • Set "API base URL" to: http://localhost:11434
   • Set "Vision model name" to the name of the model you pulled (e.g., llama3.2-vision:11b)
   • Set "Endpoint LLM model name" to the same model name (e.g., llama3.2-vision:11b) or another one. Note that it may be beneficial to use the same model as the vision model due to system resource allocation.

Verifying the Setup

To verify everything is working:

1. Click the icon in the ribbon menu to open Vision Recall
2. Click the Test Config button
3. In the modal, click the Test Connection — Retrieve Models
4. If models load in the modal below, you're good to go!

Troubleshooting

If you encounter issues:

1. Ensure Ollama is running by opening a terminal and running:

   ollama ps
   

2. Check that the base URL is correct in your plugin settings

3. Verify your model was downloaded successfully:

   ollama list
   

4. If you get connection errors, ensure no firewall is blocking port 11434

For additional help, visit the Ollama documentation or the plugin's GitHub issues page.