Here is a short introduction to Ollama and, most importantly, how to use it to process hundreds and thousands of requests at once in good old Microsoft Excel.
What is Ollama
Ollama is an open platform for running language models locally. It gives a simple way to install, run and talk to many models through one command syntax and one local API. Everything runs on your own computer: your data doesn’t go to any cloud service, which matters for trade secrets, personal data and other sensitive information.
How to work with Ollama: installing and managing models
Ollama has no usual window with buttons: you manage it through the command line (Command Prompt or PowerShell). It may seem unusual, but there’s nothing to fear: the commands are short and clear. You need the command line only to install and remove models.
Installing Ollama (Windows)
Download the Ollama installer for Windows from the official website, run it and wait for the installation to finish. Then check it in PowerShell or the command prompt:
ollama –version
Adding and removing models
| Command | What it does |
|---|---|
| ollama list | shows the models installed locally |
| ollama pull llama3.2 | downloads a model |
| ollama run llama3.2 | downloads (if needed) and runs a model in chat mode |
| ollama rm llama3.2 | removes a model |
The full list of models with descriptions is on the official Ollama website, with ready-made run commands next to each model.
Checking that Ollama works
The installer usually sets Ollama up as a background service that starts automatically with the computer, so the local API is always available at http://localhost:11434. For chatting with a model, there are browser extensions; I liked Page Assist. But chat is better done with more advanced online models: what interests us is bulk processing of prompts in Excel :)
Processing requests through Ollama in Excel
The Pro version of !SEMTools has tools for working with local models right in Excel, including through Ollama. You only need to choose a model from the list of available ones (it’s assumed you’ve already found and downloaded it via the command line).
Choosing a local model
If you haven’t changed the default Ollama port (11434), everything should work like a Swiss watch. Click the corresponding menu item:
A window with the locally available models opens automatically. Choose a model and click OK. Now you can process requests in bulk, and while the model generates answers, go have a cup of tea.
There are two ways to put your graphics card to work:
Method 1: a common pre-prompt for every cell
- select the cells with the source data;
- enter a system prompt, for example “Rephrase the following text:”;
- click OK and wait for the result.
Method 2: each cell is a ready prompt
Here each cell is processed as a separate request, so it should already contain enough text. You can build the prompts by joining several cells, or with the combinations generator, which combines cells with each other to create all variations of the source entities.
- select the cells with ready prompts;
- run the processing;
- click OK and wait for the result.
Examples of using local models in Excel
- product descriptions from technical specs;
- ad texts and slogans from keywords;
- analysis and categorization of customer reviews;
- content for email newsletters;
- translation and rewriting of texts with sensitive data.
I’m sure you’ll come up with other tasks. Just remember: a local model is only as good as your hardware lets it be, and smaller models make more mistakes, so check the results.
Run a local AI model through Ollama and process big tables right in Excel, with your data never leaving your computer.
!SEMTools solves this and hundreds of other tasks in a couple of clicks!
Summary
I hope you found what you were looking for on this page and were able to connect Ollama to Excel. Install Ollama, download a model with one command, choose it in !SEMTools and process whole columns of prompts locally. For cloud models, see the guides for ChatGPT and OpenRouter, and the general article on AI in Excel.