Customized provider configuration
Set a non-default API key, endpoint, or other provider option for a Mistral integration.
AI Tools · Function
createMistral
createMistral creates a configurable Mistral provider for the AI SDK. Configure the provider, then call it with a Mistral model identifier to obtain a model.
How it works
createMistral builds a Mistral provider instance using optional configuration such as an API key. The returned provider is callable: pass a Mistral model identifier to obtain a language model that can be used with AI SDK functions.
Syntax
createMistral(options?)Parameters
| Name | Type | Description |
|---|---|---|
options | MistralProviderSettings | Provider configuration settings, including options such as a custom API key or base URL. |
Returns
A configured Mistral provider instance that can create language, embedding, speech, and transcription models.
Key points
- Import createMistral from the @ai-sdk/mistral package.
- The factory accepts settings for a customized provider instance.
- Assign the configured factory result to a provider variable, commonly named mistral.
Common mistakes
- Use createMistral for customized configuration rather than the default mistral instance.
Example: Create a customized Mistral provider
import { createMistral } from '@ai-sdk/mistral';
const mistral = createMistral({
apiKey: process.env.MISTRAL_API_KEY,
});
const model = mistral('mistral-large-latest');
console.log(model.modelId);Creates a customized Mistral provider and selects a model from it.
When to use it
Language-model access
Create Mistral language models for text generation through the AI SDK provider interface.
Multiple configurations
Configure one provider instance separately from the package's default mistral export.
Interview questions
QWhat problem does createMistral solve?
createMistral is the factory for a customized Mistral provider. Import it from @ai-sdk/mistral, pass provider settings, and call the resulting provider with a model identifier.
Follow-up: How would you use the default configuration? Import mistral from @ai-sdk/mistral and call it with the desired model identifier.
QHow do you obtain a model after creating the provider?
The provider package is @ai-sdk/mistral. The factory returns a provider instance, which can then create models for Mistral APIs.
Follow-up: What should be passed to the provider instance? A supported Mistral model identifier, such as mistral-large-latest.
QWhen would you choose createMistral instead of mistral?
Use createMistral when provider settings must be customized; otherwise, the package also exposes a default mistral provider instance. Both are imported from @ai-sdk/mistral.
Follow-up: Why use the provider abstraction? It lets AI SDK code work through a standardized model interface across providers.
Related
mistral@ai-sdk/mistralmistral-large-latestmistral-small-latestpixtral-large-latest