Supercharge Brave Leo: Get Faster, More Private AI Help Right on Your Computer!

You probably use AI tools like chatbots for quick answers or to summarize web pages. Brave Browser has a super handy AI helper called Leo. But usually, when you ask Leo a question, it sends your request to a computer on the internet that Brave manages. What if you could get Leo's smart help, but have the AI brain run *directly on your own computer*? This means your questions never leave your device, and answers appear much faster! --- ### The Problem with Relying on Online AI * **Privacy Worries:** When you use online AI, your questions and the web page content you're asking about get sent to someone else's servers. You might not want sensitive information leaving your computer. * **Slight Delays:** There's always a tiny wait as your questions travel over the internet and answers come back. * **Hidden Costs/Limits:** Some online AIs might eventually have usage limits or subscription fees. * **No Offline Use:** You need an active internet connection to use them. --- ### The Solution: Brave Leo + Ollama (AI on Your PC!) There's a fantastic, free program called **Ollama** that makes running powerful AI models (like different "AI brains") right on your own computer incredibly easy. Brave has smartly added a feature called "Bring Your Own Model" (BYOM) that lets Leo "talk" to Ollama. **Here's why this setup is a game-changer for you:** * **Your Info Stays Private:** This is the biggest win! Your questions, thoughts, and what you're reading on a web page *never leave your computer*. Complete data privacy for sensitive topics. * **Super Fast Answers:** Because the AI brain is running right there on your machine, answers pop up almost instantly. No waiting for data to travel across the internet. * **Completely Free to Use:** Once you download Ollama and your chosen AI model, there are no ongoing costs or hidden fees for using it. * **Works Even Without Internet:** Once the AI model is downloaded to your computer, you can use Leo (with Ollama) even if your internet goes out! * **More Choices for AI Brains:** Ollama lets you pick and download different AI models from a big library. So, you're not stuck with just one type of AI; you can choose the "brain" that works best for different tasks. --- ### Exact How-To: Get Leo Using Your Local AI 1. **Get Ollama (Your Local AI Helper):** * Go to **ollama.com** and download the Ollama program for your computer (Windows, macOS, or Linux). Install it like any other app. * Once installed, open your computer's "Terminal" or "Command Prompt" (you can usually find it by searching for "Terminal" or "CMD"). * Type `ollama run llama3` (or `ollama run mistral`, `phi3`, etc., for a different AI brain) and press Enter. This will download the AI brain to your computer. Keep Ollama running in the background whenever you want to use Leo with your local AI. 2. **Configure Brave Leo to use Ollama:** * Open your **Brave Browser**. * Click on the **three-line Menu icon** (☰) in the top-right corner of the browser. * Go to **Settings**. * In the left sidebar, click on **Leo**. * Scroll down to the **"Bring your own model"** section. * Click on **"Add new model."** * Fill in the details: * **Label:** Give it a simple name you'll remember (e.g., "My Local Llama3," "Fast Mistral"). * **Model request name:** This must be the *exact* name of the model you downloaded with Ollama (e.g., `llama3`, `mistral`). * **API Key:** Leave this blank for Ollama. * **Server endpoint:** Type `http://localhost:11434/v1/chat/completions` (This is where Ollama listens for requests). * Click **"Add model."** 3. **Start Using Your Local AI in Leo:** * Now, open the **Leo AI sidebar** in Brave (click the Leo icon or `Ctrl+Shift+L` / `Cmd+Shift+L`). * At the top of the Leo chat panel, click on the **model selection dropdown menu** (it might say "Brave Leo" by default). * Choose the custom model you just added (e.g., "My Local Llama3"). That's it! Now, when you ask Leo questions, you're getting smart AI help directly from your computer, ensuring your information stays private and you get lightning-fast responses.