The Absolute Beginner's Guide to Using a Proxy with Janitor Part One
Hello! Welcome. You've heard people talking about "proxies" and "new models" and want to see what all the fuss is about. The good news is that it's surprisingly simple.
This guide will walk you through it, step-by-step. We'll get you from zero to chatting with a powerful new AI model in just a few minutes.
Our Plan:
- Part 1: The "Why?" - We'll quickly explain what a proxy is and why you want one. (No technical jargon, promise!)
- Part 2: The "How?" - A step-by-step walkthrough to get your very first proxy running for free.
- Part 3: The "What's Next?" - Now that you're a pro, we'll explore other models and options.
- Part 4: The "Oh No!" - A simple guide to fixing things if they go wrong.
Part 1: The "Why?" - Understanding the Basics
So, what is Janitor.AI's default model?
JanitorAI has its own, free AI model called JanitorLLM (or JLLM). Think of it as the "brain" that powers the characters you chat with. It's good, but there are many other, very different brains out there.
And what is a "Model"?
A "Model" (or Large Language Model/LLM) is just that: the AI brain. Different models have different personalities. Some are better at creative writing, some are smarter, some have better memory, and some are less filtered. Using a new model can completely change your roleplaying experience for the better.
What is a "Proxy"?
A proxy is simply a bridge. It's the service that connects Janitor.AI to these powerful new models.
To use this bridge, you need two things: a password and an address.
- API Key: This is your unique password. It proves to the proxy service that you're allowed to use their models.
- API URL: This is the specific address of the proxy service.
That's it! All we're going to do is get a password and an address, and tell Janitor.AI where to find them.
Part 2: The "How?" - Your First Proxy Setup (with OpenRouter)
We'll use a popular service called OpenRouter for this first setup. It's reliable and has a great free option that's perfect for starting out. You get 50 free messages every day.
Let's do this in two stages:
- Get your "stuff" (API Key) from OpenRouter.
- Put your "stuff" into JanitorAI.
Stage 1: Getting Your API Key from OpenRouter
- Go to OpenRouter: Open a new tab and go to
https://openrouter.ai/ - Create an Account: Click "Sign In" in the top right, then choose to sign up with Google or another method. It's free.
- Go to Your Keys: Once you're signed in, click your profile icon in the top right, and then click "Keys" from the dropdown menu.
- Create a Key: Click the blue "+ Create Key" button. Name it whatever you want (e.g., "JanitorAI Key"). Leave the other fields blank. Click "Create".
- COPY YOUR KEY! Your key will appear. It will start with
sk-or-v1-.... This is the only time you will ever see the full key. Copy it immediately and paste it into a safe place, like a notepad app.
That's it for Stage 1! You now have your password. Let's go use it.
Stage 2: Setting up JanitorAI
- Open Janitor.AI: Go back to your Janitor.AI tab and open a chat with any character.
- Open API Settings: At the top of the chat, click the three dots (...) and select "API Settings".
- Select Your API: In the "AI Provider" dropdown, select "OpenRouter".
- Paste Your Key: In the "OpenRouter API Key" box, paste the key you just saved from their website.
- Choose a Model: We need to tell Janitor which brain to use. For your first time, let's use a great, reliable one. In the "Model" dropdown, find or type in:
deepseek/deepseek-chat-v3-0324:free - Important: Make sure you select the one that ends in
:free! - Check Settings & Save: Your setup should now look like this. Click the "Save Settings" button at the bottom.
- REFRESH THE PAGE! This is a crucial final step. Hit the refresh button on your browser.
Congratulations! You did it! You can now chat with the bot using a brand new, powerful AI. Enjoy your 50 free daily messages!
Part 3: The "What's Next?" - Exploring Your Options
Now that you've got the hang of it, you can explore.
Trying Different Models
You can change the model anytime in the API Settings. Here are a few popular free ones on OpenRouter to try (just type their name in the "Model" box):
deepseek/deepseek-chat-v3-0324:free: The one we just used. A fantastic all-rounder. Softer and more agreeable.deepseek/deepseek-r1-0528:free: A "reasoning" model. It's smarter, better at following instructions, and can be more intense or "mean" if the character calls for it. It will show its thoughts in<think>tags before replying.microsoft/mai-ds-r1:free: A tamer, more "SFW" version of R1. Good for softer stories.
Getting More Free Messages on OpenRouter
If you find yourself running out of the 50 free messages, OpenRouter has a great offer. If you add $10 of credits to your account once, your daily limit for free models is permanently boosted to 1,000 messages per day. You don't have to spend the credits; just having them on your account unlocks the higher limit forever.
What about a Higher Context Limit?
In your Chat Settings, you'll see "Context Size." This is the AI's memory. While bigger seems better, most models work best with a context of 16,384 (16k). Any higher can actually make them slower and more forgetful.
Part 4: The "Oh No!" - Simple Troubleshooting
Getting an error? Don't panic. 99% of issues are one of these simple things.
Before you do anything else, try these three steps:
- Re-save your settings: Go back to API Settings and hit "Save Settings" again.
- Refresh your browser tab: A full page refresh solves most problems.
- Check your key: Make sure you copied the entire key from OpenRouter, with no extra spaces.
And if you are still getting error messages, we've got you covered! Simply tell the chatbot what you're seeing (an error 401 for example!) and it'll sort you out. Or you can check out this article on Troubleshooting Deepseek via OpenRouter! Beyond that, you can join our Reddit or Discord server and ask in there!
Next Up: The Absolute Beginner's Guide to Using a Proxy with Janitor Part Two
Updated on: 02/08/2025
Thank you!