[Feature] Native Mistral AI Provider Support

[sabitvw-code]

MaxKB Version

MaxKB version: v2.8.1 (self-hosted Docker on Windows 11)

Please describe your needs or suggestions for improvements

Feature Request: Add Mistral AI as a native model provider

Problem

MaxKB supports OpenAI, DeepSeek, Ollama, and many other providers, but Mistral AI — one of Europe's leading AI labs — is missing from the built-in provider list.

Mistral API is fully OpenAI-compatible (/v1/chat/completions, /v1/models) and uses Bearer token authentication — identical to other already-supported providers.

Why Mistral?

• Mistral Large ranks among top models for RAG/knowledge-base tasks
• Fully OpenAI-compatible API → minimal implementation effort
• Already supported in Dify as a native provider
• Growing European user base

Current Workarounds (Broken)

Users try:

1. Adding as "OpenAI" provider with custom base URL → model list works but verification/runtime fails
2. Direct database insertion with RSA-encrypted credentials → unstable, breaks after UI changes
3. Using Ollama's mistral:7b instead → limited to local 7B model, no access to Mistral Large

Suggested Implementation

• Add model_mistral_provider following the same pattern as model_deepseek_provider and model_openai_provider
• Use LangChain's langchain-mistralai package or reuse existing OpenAI-compatible wrapper
• Base URL: https://api.mistral.ai/v1
• Auth: Authorization: Bearer <API_KEY>
• Supported models: mistral-large-latest, mistral-small-latest, pixtral-large-latest

Context

We recently submitted a similar feature request to DeepChat — it was implemented and merged within 24 hours (PR #1598). MaxKB is a critical part of our AI stack for automotive diagnostics and Mistral support would be a game-changer.

MaxKB version: v2.8.1 (Docker)
Deployment: Docker on Windows 11

Please describe the solution you suggest

No response

Additional Information

No response