Skip to content
NEW SELECTThe new Itzi Select plans are here
Itzi
Open app
All connection guides

Connect JanitorAI to Itzi

Add Itzi as a proxy in your chat settings, then enter your key and model ID.

Create your Itzi key

Open the Itzi API dashboard and sign in. Under API keys, enter a name for this app and click Create. Copy the secret immediately: it is shown only once. Keep a separate key for each app and paste it only into its API key field.

Choose a model your account can access. Requests use your Itzi balance or subscription according to the key settings. The example keys in these screenshots do not work.

Open API dashboard

Connection details

Proxy URL

https://itzi.app/v1/chat/completions

This field expects the full generation URL, including /chat/completions.

API key
Your full Itzi key, starting with itzi_sk_.
Model
An exact model ID from your API dashboard or /v1/models.

Set up the connection

  1. Open the chat settings

    Sign in to JanitorAI and open a chat that allows external proxies. Click the three-line menu at the top right of the chat, then choose Settings. The settings panel opens on the right.

    JanitorAI chat menu with Settings as the first option below the three-line menu button.
    Open the chat menu and choose Settings. Open the image to view it at full size.

  2. Choose the proxy provider

    In the Settings tab, set Provider to Proxy. Under Proxy configurations, click + New to open the New Configuration form. If the character does not allow proxies, use one that does.

    JanitorAI settings with Provider set to Proxy, an empty Proxy configurations list, + New buttons and Save at the bottom.
    Set Provider to Proxy, then use + New to add Itzi. Open the image to view it at full size.

  3. Enter the full generation URL

    Set Name to Itzi. Paste the full URL from Connection details into Proxy URL. It already includes /v1/chat/completions, so do not append the path again. The screenshot shows an empty form; replace the placeholder URL with the Itzi URL.

    JanitorAI New Configuration form with Name, Proxy URL, API key, Model and Custom prompt fields, followed by Add and the main Save button.
    The fields are empty here. Enter your Itzi details, click Add, then Save. Open the image to view it at full size.

  4. Add the key and model

    Paste your full Itzi key into API key. Itzi requires a key even though the field is marked optional. Set Model to an exact model ID from the Itzi API dashboard, not its display name. Leave Custom prompt empty for the first test.

  5. Add, save and refresh

    Click Add in the New Configuration form, select Itzi if needed, then click Save at the bottom of the Settings panel. Refresh the page after adding or changing the proxy URL, as JanitorAI requests. Send a short message to test the connection; this uses your Itzi balance or subscription allowance.

If something does not work

Save is disabled while the form is open

Finish the new configuration with Add before using the main Save button. If JanitorAI says Save or cancel proxy config first, the New Configuration form is still open.

Check the exact endpoint field

The Proxy URL described here uses the complete generation URL. If your version instead explicitly asks for a Base URL and says it appends /chat/completions, use https://itzi.app/v1 there.

A blank reply needs a smaller test

Try a short prompt with default generation settings and a model available to your key. If it still fails, check Itzi Status and the same model in Itzi Chat before retrying the proxy.

Invalid key or 401

Copy the entire Itzi key without spaces. If it was revoked or lost, create a replacement in the API dashboard. An upstream provider key cannot replace an Itzi key.

Payment or access error

Check your balance, subscription allowance and the API key's subscription settings. A model shown in the public catalog may require access your key does not have.

Model not found or an empty model list

Refresh the model list using your key. Copy the exact ID from the API dashboard, then confirm the endpoint and key were saved before model discovery.

Too many requests or 429

Wait before retrying and reduce parallel requests. Chat messages and background agents share your limits. Check the error message for a rate, concurrency or subscription limit.

Long chats fail or responses are empty

Test a short conversation first. Lower the client's context and output limits, including character, lorebook and system prompts in that budget. Check Itzi Status if a short request also fails.

Check Itzi Status