xAI (Grok)

xAI (Grok) lets RealTimeX use Grok-family hosted models through xAI's API.

Open it from Settings > AI Providers > LLM, then choose xAI (Grok) or xAI, depending on the current build label.

What you need

Before RealTimeX can use xAI:

• you need a valid xAI API key
• the RealTimeX instance must be able to reach the xAI API
• you should know which Grok model family you want to use for chat

Current setup flow

1. Open Settings > AI Providers > LLM.
2. Choose xAI (Grok).
3. Paste your xAI API key.
4. Wait for the model list to load.
5. Select the Grok model you want to use.
6. Save the provider settings.
7. Test a normal chat.

Model selection behavior

The current xAI setup uses live model discovery.

Important behavior:

• the model dropdown stays unavailable until a usable API key is entered
• if the key is valid, RealTimeX loads the available Grok models for that account
• if the model list changes on the xAI side, reopening or refreshing the selector picks up the updated catalog

When to use xAI

• Use xAI (Grok) when your team standardizes on Grok models.
• Use it when you want hosted inference without managing a local runtime.
• Use it when you want provider-native access rather than routing through a compatibility layer.

Troubleshooting

No models appear after entering the key

Check that:

• the API key is correct
• the RealTimeX host can reach xAI
• your xAI account has access to the model catalog you expect

I want to route Grok through another gateway

If you do not want the vendor-native integration, use OpenAI (generic) or another compatibility layer instead of the dedicated xAI provider.