Directus
Using mittwald AI Hosting with Directus
Directus v11.15 and later includes built-in AI assistant capabilities with support for OpenAI-compatible providers. Since mittwald AI Hosting provides an OpenAI-compatible API, you can use it directly with Directus without installing any additional extensions.
Prerequisites
Ensure you are running Directus version 11.15.0 or later, as this version introduced multi-provider AI support including OpenAI-compatible endpoints.
If you don't have an API key yet, follow the mittwald AI Hosting access guide to generate one through your mStudio dashboard.
Configuration
The AI assistant in Directus is configured through the admin settings interface. Navigate to Settings → AI in your Directus admin panel to access the OpenAI-compatible provider configuration.
Step 1: Configure the provider
In the AI settings, configure the OpenAI-compatible provider with the following values:
- Base URL: Enter
https://llm.aihosting.mittwald.de/v1 - API Key: Enter your mittwald AI Hosting API key
- Model: Enter the model identifier you want to use, such as
ministral-3-14b-instruct-2512
The system automatically appends /chat/completions to the base URL when making requests, so do not include this path segment in the base URL field.
Step 2: Enable the AI assistant
After configuring the provider, ensure the AI assistant is enabled. If AI features are disabled at the system level through the AI_ENABLED environment variable, the AI assistant sidebar will be hidden and the settings form will display a disabled notice.
Available models
You can use any chat-capable model from mittwald AI Hosting with Directus. Recommended models include:
ministral-3-14b-instruct-2512— General-purpose model with vision supportdevstral-small-2-24b-instruct-2512— Optimized for code-related tasksgpt-oss-120b— Large model for complex reasoning tasks
For a complete list of available models and their capabilities, see the models overview.
Usage limits
The mittwald AI Hosting service has usage limits based on your account tier. For details on rate limits and quotas, refer to the terms of use.