Ollama

This tutorial will guide you step-by-step on how to connect to a local model from Ollama.

Prerequisites

• Considering that local models are running, it is recommended to use a computer with good performance.

Steps

• Download and install the Ollama application from the Ollama official website.

• After installation, open the Ollama application and wait for it to initialize. Then, open a command-line tool (such as PowerShell or CMD) and enter the following command to verify if the Ollama CLI is available:

ollama -v

• If the command line outputs the version number of Ollama, the installation is successful. Otherwise, please refer to the Ollama official documentation. Next, you can choose and download a local model. Search for models supported by Ollama here. For example, to download the gemma3:12b model:

ollama pull gemma3:12b

• After the download is complete, enter ollama list in the terminal to ensure that the downloaded model is in the results.
• In Everywhere's Custom Assistant, switch to Advanced Mode:
  • Set URL to http://127.0.0.1:11434 (Ollama's default address)
  • Set Model ID to gemma3:12b (or the name of another model you downloaded).

• Once the setup is complete, you can start using it.

FAQ

Chat message displays ModelDoesNotSupportTools

If you encounter a chat message displaying ModelDoesNotSupportTools when using an Ollama model, it usually means that the Ollama model you are using does not support the tool-calling feature. You can try switching to another model that supports tool calling, or disable the tool-calling feature in Everywhere to avoid this error.