Anvia

Mistral ​

@anvia/mistral connects Mistral to Anvia's provider-neutral model interfaces. The same client can create models for chat completions, embeddings, and OCR while the rest of the application stays on Anvia APIs.

ts
import { MistralClient } from '@anvia/mistral'

export const mistral = new MistralClient({
  apiKey: process.env.MISTRAL_API_KEY!,
})

What the client provides ​

  • completionModel(...) for agents, direct completions, tools, and structured output
  • embeddingModel(...) for retrieval and semantic-search vectors
  • ocrModel(...) for extracting Markdown and page data from documents
  • listModels() for administrative model inventory

Start with Setup, then open the guide for the capability you need:

Completion boundary ​

The current completion adapter declares support for streaming, tools, tool choice, and output schemas. It does not declare chat image input, chat document input, normalized reasoning content, or provider-executed tools.

OCR is therefore a separate document-processing path. Extract a document with ocrModel(...), then decide whether its text should enter a prompt or a knowledge-ingestion pipeline.

Capability declarations describe the adapter contract, not every upstream model ID. Test the exact model and request shape used by the deployment.

Application responsibilities ​

The provider package maps requests, normalizes streams and usage, creates embeddings, performs OCR, and lists models. Your application still owns:

  • instructions and model selection
  • tool authorization and side effects
  • tenant and document access
  • retries, timeouts, and fallback policy
  • retention and observability

Review the production checklist before shipping.

Built for Anvia.