This archived newsletter issue covering local LLMs for freelance writers was originally sent to subscribers on June 23, 2026.
When you think of large language models (LLMs), you probably think of the big cloud-based AI tools like ChatGPT, Claude, and Gemini.
But did you know there’s something called “local LLMs” that offer complete privacy, local runs (meaning no use of datacenter resources), and no chance of your inputs being used for training data?
While it’s true local LLMs (like Google’s Gemma 4) aren’t as powerful as the frontier models you might have access to (if you pay for premium AI tools), they can handle more than you might realize.
What are Local LLMs?
AI models (think GPT 5.5 and Claude Opus 4.8 as examples) generally run through cloud services. All the processing happens in datacenters, and it can be very resource-intensive.
Local LLMs, on the other hand, are designed to be smaller. You download the full model itself onto your local computer (or even phone in some cases), and everything runs on that individual machine.
Nothing you enter is sent to the cloud. Nothing is processed by a third-party company. No one else can access that data, your chats, or anything you feed into them.
Why Local LLMs Can be Great for Freelancers
That privacy makes local LLMs ideal for things like private data analysis (think financial or health-related info you wouldn’t want anyone accessing). They’re also a good option for processing business data you wouldn’t want to upload.
For example, I don’t consider it ethical to upload client information to a cloud-based AI tool. No one has the right to risk that data being used for AI training other than the client themselves.
With local LLMs, that’s not a concern. Everything stays private on your own system.
For instance, I have a client that provides extremely detailed project briefs. It’s easy to miss something minor.
So in our last project, I took my project draft and that brief into Gemma 4, and it was able to quickly let me know what was covered and what was missing so I could make final edits faster.
While I’d never suggest using AI to generate drafts you plan to deliver to clients, these local models can help with ideation, outlining, summarizing research material, comparing your drafts to custom style guidelines, and much more.
Remember, “AI tools” don’t only cover content generation. There’s also agentic AI, data analysis, and automation that can be very useful in managing a freelance, or any, business.
Local LLMs give you the option to experiment with some of these things without worrying about who might access your, or your clients’, data. Oh, and it’s completely free.
—
If you read the latest post on the blog about Obsidian, and you’re interested in the idea of using a local LLM, keep an eye on future updates where we’ll talk about building out an “AI operating system” in Obsidian that your local LLM and other desktop AI apps can use to maintain consistency between AI agents you build and the tools you use.
Would you like to receive content like this in your inbox? Get subscribers-only content, and early content before it’s archived on the website, by subscribing to the Freelance Writing Pros newsletter.