Using Continue Plugin With Ollama as GitHub Copilot Alternatives

I recently stumbled upon a game-changing combo: Ollama and the Continue plugin. Trust me, this open-source duo is about to revolutionize the way you code.

Why Continue and Ollama Are My New Best Friends

Ollama an open-source platform that lets you run language models locally. It's lightweight, efficient, and supports a ton of models. I've been playing around with Llama 3 and StarCoder, and I'm pretty impressed.

Now, Continue is where the magic really happens. It's a plugin for VS Code and JetBrains that works seamlessly with Ollama. The interface is super intuitive, and the features? Let's just say my productivity has skyrocketed:

• Smart Code Completion: It's like it reads my mind, offering spot-on suggestions as I type.
• Refactoring Wizardry: Cleaning up my code has never been this easy or fun.
• Code Understanding: I can ask questions about my codebase right in my editor, and boom – clear, concise answers.

The best part? My wallet's definitely thanking me. Plus, I sleep better at night knowing all my code stays right here on my laptop. And man, being able to tinker with every little setting? It's like I've got my own personal coding genie. Seriously, it feels like I've unlocked some secret dev superpower that's perfectly tuned to how my brain works.

Setting Up Your Own AI Coding Sidekick

Alright, let's get you set up with this awesome combo. Don't worry, it's not as daunting as it might sound!

Step 1: Get Ollama Up and Running

Head over to the Ollama website (https://ollama.com) and grab the version that matches your OS. The installation is pretty straightforward – just follow their instructions, and you'll be good to go.

Step 2: Time to Pick Your Models

This is where it gets fun. Ollama supports a bunch of models, but for coding, I've found these to be real winners:

• StarCoder 2 3b: Great all-rounder for everyday coding tasks.
• DeepSeek Coder 6.7b-base: When you need some extra firepower for complex stuff.
• DeepSeek Coder 1.3b-base: A lighter option that still packs a punch.

To get these models, just use the ollama pull command. For example:

ollama pull starcoder2:3b

Choose the one that fits your style – or hey, why not grab all three?

Step 3: Continue Installation

Next up, we need Continue. Pop over to https://continue.dev and download the plugin for your favorite IDE (VS Code or JetBrains). Installation is a breeze – just follow their guide.

Step 4: Getting Continue to Play Nice with Ollama

This is where we connect all the dots. Open up your Continue settings (it's in the preferences menu of your IDE), and let's make some tweaks:

• Set the "Provider" to ollama.
• Choose your model (like starcoder2:3b) in the "Model" field.
• If you're running everything locally, you can leave the "API Base" blank. But if you're fancy and hosting Ollama on a separate machine, pop in the URL here.

There are tons of other settings you can play with to really make it your own. I spent a good hour tweaking things like completion options and tab autocomplete settings.

Step 5: Start Coding

That's it! Fire up your IDE, start coding, and your new AI assistant will start does its thing.

Wrapping Up

Look, I know setting up new tools can be a pain, but hear me out. Ollama and Continue are seriously game-changers. Yeah, it might take you an afternoon to get everything running smooth, but oh boy, is it worth it.

I've been tinkering with this setup for a few weeks now, and I gotta tell you, I'm kicking myself for not doing it sooner. My code's actually starting to look halfway decent, and I'm knocking out tasks in half the time. The other day, I tackled this gnarly bug that would've had me pulling my hair out before, and I actually had fun doing it. Can you believe that?

So do yourself a favor and give it a go. Fair warning though - you might get a little addicted. Don't blame me if you start looking for excuses to code!

P.S. Got questions? Drop them in the comments. I'd love to hear about your experiences with this setup!