Skip to content

FAQ

Is DXPR AI Provider free to use?

Yes. Every DXPR account includes free monthly AI credits: 10,000 credits for personal accounts and 10,000 plus 1,000 per user for paid plans. One credit equals one generated word; one generated image equals 1,000 credits. HTML markup in text responses does not count toward credit usage, only readable content. Additional credits are available for purchase at dxpr.com/pricing.

Which AI models does DXPR AI Provider give access to?

The provider connects to the DXAI Kavya platform, which routes requests across five AI providers: OpenAI, Anthropic Claude, Google Gemini, XAI, and MistralAI. Three model variants are available:

  • kavya-m1: defaults to Anthropic Claude with failover through OpenAI, Mistral, Gemini, and XAI
  • kavya-m1-eu: EU data residency via MistralAI (Mistral Large with Mistral Small fallback)
  • kavya-m1-fast: low latency via Gemini Flash Lite with XAI Grok fallback

For image generation, the kavya-image model handles both text-to-image and image-to-image operations.

How does multi-provider failover work?

When you use kavya-m1, the Kavya platform tries your highest-priority enabled provider first. If that provider fails or times out, it automatically retries with the next provider in your configured order. This continues until a provider succeeds or all enabled providers have been tried.

You configure provider priority through a drag-and-drop table at /admin/config/ai/providers/dxpr. Each provider can be enabled, disabled, and reordered. See the multi-provider failover guide for details.

How does em-dash removal work?

AI models frequently overuse em-dashes, the most recognisable marker of AI-generated text. DXAI's post-processing engine detects em-dashes in generated content and replaces them with natural punctuation. Four intensity levels are available: Off, Low, Medium, and High (the default).

The em_dash_mode value is sent with every chat request to the Kavya API, which applies the replacement server-side. On multilingual sites, each language can have its own intensity level. See the em-dash post-processing guide for details.

Does DXPR AI Provider support EU data compliance?

Yes. The kavya-m1-eu model routes through the Kavya server in the Netherlands (backup in Germany) and MistralAI servers in France, keeping all processing within European boundaries. Select this model in the provider configuration to enforce EU-only routing. It is designed for organisations subject to GDPR or other EU data residency requirements.

Can I use my own API keys from OpenAI or Anthropic instead?

No. DXPR AI Provider connects exclusively to the DXAI Kavya platform using a DXPR API key. It does not accept API keys from individual AI providers.

If you want to use your own OpenAI or Anthropic API keys directly, install the respective provider modules (such as the OpenAI provider or Anthropic provider) from the Drupal AI module ecosystem. These can run alongside DXPR AI Provider; the AI module's provider selection logic will route requests to the appropriate provider based on your configuration.

I set up translation but some fields are not being translated. Why?

The most common cause is that Content Translation is enabled for the content type but individual fields are not ticked. Navigate to /admin/config/regional/content-language, expand the content type, and make sure each field you want to translate has its checkbox enabled. The title field is separate from the body field; both must be ticked individually. See the translation setup guide for the full walkthrough.

Does DXPR AI Provider support streaming?

Yes. The module supports streamed chat responses via server-sent events (SSE). When a consuming module requests streaming (as AI Assistants API does for its chat interface), DXPR AI Provider opens a streamed connection to the Kavya API and yields response chunks as they arrive. Streaming preserves custom metadata fields (word count, original model) that the standard OpenAI PHP client would strip. All three model variants (kavya-m1, kavya-m1-eu, kavya-m1-fast) support streaming.

What happens when I run out of credits?

When your account's credit balance is exhausted, the Kavya API returns a rate-limit error. The module surfaces this as a clear error message indicating that additional credits are needed rather than a generic failure.

Free credits reset monthly. Additional credits are available for purchase at dxpr.com/pricing. You can monitor your current credit balance and monthly usage in the AI Usage Tracking section of the DXPR AI Provider settings page.