Mistral MCP Server

MCP server exposing Mistral AI capabilities over MCP: chat, embeddings, FIM, vision, OCR, audio, agents, moderation, classification, files, batch

Remote serverstreamable-http

What is the Mistral MCP MCP server?

Connect Mistral MCP to Claude, Cursor or any other MCP client and it stops being a tab you switch to. MCP server exposing Mistral AI capabilities over MCP: chat, embeddings, FIM, vision, OCR, audio, agents, moderation, classification, files, batch, workflows, conversations, RAG libraries, sampling, prompts, resources, and Streamable HTTP. The mistral mcp mcp server is what makes that connection.

What the server does

mistral-mcp exposes the full Mistral AI API as a set of MCP tools, resources, and prompts. An MCP client (Claude Code, Cursor, etc.) can call mistral_ocr to extract text from a PDF, voxtral_transcribe to transcribe a meeting recording, or workflow_execute to start a durable multi-step process — all without leaving the agent loop.

  • self-hosted MCP server, no mandatory SaaS proxy
  • bring-your-own Mistral API key (BYOK) — Mistral states API data is not used to train its models
  • lean core profile and focused metier-docs profile to limit tool exposure
  • process_document cache configurable per-call and via MISTRAL_MCP_CACHE_DIR
  • ID document cache bypass enabled by default, even when kind:"auto" resolves to id_document
  • Streamable HTTP + bearer auth path for controlled / on-prem deployments

Available tools

The toolset is worth reading before you wire it up, because it tells you what the integration is really for:

  • mistral_chat — Chat completion. Supports all Mistral models, response_format, reasoning_effort for Magistral
  • mistral_vision — Multimodal chat with images (URL or base64)
  • mistral_ocr — Document AI — extract text, bbox, and JSON annotations from PDFs/images. Pass includeBlocks: true for OCR 4 paragraph-level blocks
  • codestral_fim — Fill-in-the-middle code completion (Codestral model)
  • voxtral_transcribe — Audio → text. Pass diarize: true for speaker separation
  • workflow_execute — Start a Mistral Workflow (Temporal-backed durable execution)
  • workflow_status — Poll a running workflow — returns RUNNING \
  • workflow_interact — Signal / query a running workflow. Used for human-in-the-loop checkpoints
  • connectors_list — Discover Mistral Connectors (MCP/HTTP integrations) visible to the caller
  • connectors_get — Fetch one connector's public metadata (never credentials)
  • connectors_list_tools — List the MCP tools a connector exposes, with their input schema
  • connectors_call_tool — Invoke a connector's tool — real MCP CallToolResult passthrough

Credentials and setup notes

Configuration is passed through the environment: MISTRAL_API_KEY, MCP_HTTP_HOST, MCP_HTTP_PATH, MCP_HTTP_TOKEN. Treat anything key-shaped as a real credential — scope it to the minimum the server needs, and rotate it if it ever lands in a shared config.

Installation

Because this one is hosted, setup is mostly authentication — you point your client at the endpoint and approve access. Nothing runs on your machine, so there is no runtime to keep patched.

Where it fits

This sits in the team communication group, where several servers overlap in what they claim to do but differ sharply once you actually set them up. Mistral MCP's toolset — mistral_chat, mistral_vision, mistral_ocr and 11 more — is a fair guide to whether it matches your workflow. It is maintained by swih; worth a glance at recent repository activity before you build anything load-bearing on it.

This entry was verified against Mistral MCP's own documentation before publication; SyncDev keeps the directory reviewed rather than auto-generated.

Worth knowing first

  • Your data travels to the provider's service, so the usual questions apply about what you send and what they retain.
  • With 14 tools registered it takes up a noticeable share of the context window; turn it off in projects that never touch Mistral MCP.
  • Missing credentials fail quietly in some clients — if no tools show up, check the environment block first.
  • Keep per-call confirmation enabled while you learn its behaviour; it is the cheapest safeguard you have.

Available tools

ToolWhat it does
mistral_chatChat completion. Supports all Mistral models, response_format, reasoning_effort for Magistral.
mistral_visionMultimodal chat with images (URL or base64).
mistral_ocrDocument AI — extract text, bbox, and JSON annotations from PDFs/images. Pass includeBlocks: true for OCR 4 paragraph-level blocks (text/title/table/image/equation/... with bounding boxes).
codestral_fimFill-in-the-middle code completion (Codestral model).
voxtral_transcribeAudio → text. Pass diarize: true for speaker separation.
workflow_executeStart a Mistral Workflow (Temporal-backed durable execution).
workflow_statusPoll a running workflow — returns RUNNING \
workflow_interactSignal / query a running workflow. Used for human-in-the-loop checkpoints.
connectors_listDiscover Mistral Connectors (MCP/HTTP integrations) visible to the caller.
connectors_getFetch one connector's public metadata (never credentials).
connectors_list_toolsList the MCP tools a connector exposes, with their input schema.
connectors_call_toolInvoke a connector's tool — real MCP CallToolResult passthrough.
GroupTools
Generationmistral_chat_stream, mistral_embed, mistral_tool_call

How to install the Mistral MCP MCP server

{
  "mcpServers": {
    "mistral": {
      "command": "npx",
      "args": ["-y", "mistral-mcp"],
      "env": {
        "MISTRAL_API_KEY": "your-value",
        "MCP_HTTP_HOST": "your-value",
        "MCP_HTTP_PATH": "your-value",
        "MCP_HTTP_TOKEN": "your-value"
      }
    }
  }
}

Add to claude_desktop_config.json, then restart Claude Desktop.

Configuration

VariableDescriptionRequired
MISTRAL_API_KEYCredential the server authenticates with.Yes
MCP_HTTP_HOSTEndpoint or connection string the server talks to.Optional
MCP_HTTP_PATHFilesystem location the server is allowed to use.Optional
MCP_HTTP_TOKENCredential the server authenticates with.Yes

Example prompts to try

  • Use Mistral MCP to mistral chat.
  • Use Mistral MCP to mistral vision.
  • Use Mistral MCP to mistral ocr.

Frequently asked questions

It connects Mistral MCP to MCP-compatible AI assistants such as Claude and Cursor, exposing 14 tools (mistral_chat, mistral_vision, mistral_ocr, and more) that the assistant can call on your behalf. Instead of copying data back and forth by hand, the assistant works with Mistral MCP directly.