← All guides

Connect JanitorAI to UnsavedAi

Use UnsavedAi as the proxy API for your JanitorAI characters in about a minute.

Get your three connection values

Create an account and a key named JanitorAI in your UnsavedAi console. You need the proxy URL https://api.unsavedai.com/v1/chat/completions, that API key, and a model ID from the Models page, such as gemma-4-12b.

Connect in JanitorAI

Open a chat, open API Settings, and choose the proxy option for a custom OpenAI-compatible API. Paste the full URL including /chat/completions, type the model ID exactly, and paste your key. Save the settings and refresh the chat. Labels can change between JanitorAI versions.

Generation settings

A context size of 32,768 tokens is a good start; the limit is 40,960 including the reply. Max new tokens can be 0 to use our default, or 300 to 800 for shorter replies. A temperature between 0.8 and 1.0 suits most characters. Settings JanitorAI does not send use the model defaults.

If a message fails

A 401 error means the key is wrong or has been revoked. A 429 error means the simultaneous-request limit of your plan or the queue limit was reached; wait a moment and send again. The first reply in a long chat can take several seconds while the history is processed.

Your next step

Check available models, open your console, or read the API documentation.