How to use Proxy on JanitorAI and use Deepseek V3
Hello, here I will guide you step by step on how you can use a proxy in JanitorAI to roleplay with other LLMs
Step 1 - Choose a provider
Currently, there are several sites that offer the possibility of using a language model (LLM), but for this guide I will show you two possible choices
Here is a summary table of the current services if we want to use a free model:
| Limit | OpenRouter | ChutesAI |
|---|---|---|
| Free | 20 Requests/day | None |
| Minimum balance | 1000 Requests/Day for 10 credit balance (Yearly) / 50 Request/Day for less than 10 credit | 200 Request/Day for $5 minimum |
As you can see yes it is quite limited because it requires a lot of resources and more and more people are using them which forces limits to be imposed.
Currently, to use ChutesAI, you are required to deposit a minimum of $5 in order to use the free models.
OpenRouter
- First go on Openrouter and create a new account

- Once your account is created go to your profile and then to keys

- Then click on "Create API Keys" and add a name



After this you will see your key, copy it and keep it somewhere safe where you can find it again and never share it!! because you will only be able to see it once during creation (For example, in the Notes app of your phone)
- Then go to "Model" to see the list of available models

- In the filters set the price to 0 and in the search bar type "DeepSeek" which will show you the free DeepSeek models

- I recommend using DeepSeek V3 0324, click on it to see the details

- Then under the name you will see
deepseek/deepseek-chat-v3-0324:free
which is the model name to copy
A model is free when you see :free at the end of its model name

- Then go back to JanitorAI and open a chat
- Click on the Proxy button (or Janitor LLM) to open the settings and go to "Proxy"
- Then click on add a configuration

- Then fill in the parameters with this:
Configuration Name : Put the name of the model or anything else (For example DeepSeek V3 0324)
Model Name : That’s the name of your model on OpenRouter, in case of DeepSeek V3 0324 free put
deepseek/deepseek-chat-v3-0324:free
Proxy URL : This is the link that connects with OpenRouter so it can send the responses and generate the messages
https://openrouter.ai/api/v1/chat/completions
API Key : Here copy the API key you created in OpenRouter
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 (Use your own key, I deleted this one ;3)
Custom Prompt (Optional)
Leave it empty for now (Otherwise, search for custom prompts for the models you will use)
- Normally you should have the same thing as on the image in the settings

- Once this is done click on "Add Configuration" and there you go your Proxy is configured and you will be able to use DeepSeek for your conversations!

- Then go into the 3 bars at the top and into "Generation Settings"

- Adjust the parameters like these (For Deepseek V3 0324 only)
Temperature:
0.8
Max tokens :
0 (Unlimited)
Context Size :
Average ~16k

- Click on save and reload the page and you can start chatting with the bot!