Menu
Mattering Automations connector

The Mistral MCP for chat, OCR, and batch inference inside your automations

Connect one Mistral API key and chat completions, Codestral code completion, Mistral OCR, embeddings, moderation, and asynchronous batch jobs become steps on a visual canvas that AI agents can drive. 24 typed nodes over Mistral AI Studio.

24 ready-made nodes One credential to connect Works with Claude, ChatGPT, and Matter

What the Mistral MCP connector does

The Mistral AI connector for Mattering Automations covers the stable surface of Mistral AI Studio as 24 typed nodes: chat completions with reasoning on Magistral models, vision input, JSON mode and structured outputs, and function calling; fill-in-the-middle code completion with Codestral; running agents you built in Mistral's console; embeddings; text and conversation moderation plus your own fine-tuned classifiers; the Mistral OCR document model; fine-tuned model management; file reads; and the full batch inference lifecycle. Eight are read-only queries and sixteen are actions.

Every node speaks MCP, the open protocol for connecting AI assistants to tools. Claude, ChatGPT, and any MCP-capable assistant can call them, and so can Matter, the assistant built into the platform. A workable division of labor: your assistant plans the workflow and reads the results, while the bulk generation, document parsing, and classification run on Mistral, all on one canvas you can see and edit.

Three details are worth knowing up front. Model fields are free text with documented -latest defaults such as mistral-small-latest and codestral-latest, because Mistral ships and retires models on a published schedule and a fixed dropdown would rot; run Mistral: List Models for the live roster your key can call. Batch jobs accept inline requests, so bulk work runs end to end without uploading a file first. And streaming is stripped from every outbound body: each node returns one complete response.

One connection, three surfaces

  • Visual automations. Drag Mistral nodes onto the canvas and wire them to triggers, conditions, and other apps.
  • AI agents, yours or ours. These nodes speak MCP, the open protocol, so Claude, ChatGPT, and any MCP-capable assistant can use them, and Matter, the agent built into the platform, uses them natively.
  • The whole platform. Combine with native tasks, projects, and other connectors, no middleware in between.

Every Mistral node, ready to drop on the canvas

24 nodes, generated from the connector itself so this list always matches what ships. Expand a category to browse.

Read and query nodes 8

Mistral: Download File Content

Download a stored file's raw content: the primary way to collect batch job results. The full text comes back at data.content, and because batch outputs are JSONL, each line is also parsed into data.lines so a For Each node can loop over the individual responses.

Mistral: Get Batch Job

Poll one batch job by ID: status (QUEUED, RUNNING, SUCCESS, FAILED...), progress counters, and on completion the output_file and error_file IDs. With inline=true, jobs created from inline requests return their outputs directly in the response.

Mistral: Get File

Retrieve one file's metadata by ID: filename, size, purpose, line count, and source. Useful for checking a batch job's output_file before downloading it.

Mistral: Get File Signed URL

Get a temporary signed download URL for a stored file, valid for a configurable number of hours. Hand the URL to any downstream node or external system that needs the raw file without a Mistral credential.

Mistral: Get Model

Retrieve one model by ID: its capabilities (chat, FIM, function calling, vision, fine-tuning support), maximum context length, aliases, and deprecation date. Works for base models and your fine-tuned models alike.

Mistral: List Batch Jobs

List your batch inference jobs with paging and filters by model, status, and creation date. Each job carries progress counters and, once finished, the output_file and error_file IDs to download.

Mistral: List Files

List files stored in your Mistral workspace: fine-tuning datasets, batch inputs, batch outputs, and OCR documents, with paging and filters by purpose, type, source, and filename search.

Mistral: List Models

List every model your Mistral API key can use: base models with their capabilities (chat, function calling, vision, fine-tuning), context lengths, aliases, and deprecation dates, plus your fine-tuned models. Free to call, which makes it the credential health check for this connector.

Action nodes 16

Mistral: Agent Completion

Run a completion against a Mistral agent built in the Mistral Studio console: the agent carries its own model, instructions, and tool set, so this node only needs the agent ID and the messages. Output lands at data.choices[0].message.content. Streaming is forced off.

Mistral: Archive Fine-Tuned Model

Archive a fine-tuned model, taking it out of active use without deleting it. Fully reversible with the Unarchive node.

Mistral: Cancel Batch Job

Request cancellation of a running batch job. Requests already completed stay completed and billed; the job moves through CANCELLATION_REQUESTED to CANCELLED.

Mistral: Chat Completion

Generate a chat completion with a Mistral model: frontier reasoning with Magistral, multimodal input on vision models, JSON mode and structured outputs, and function calling. The generated text lands at data.choices[0].message.content. Streaming is not supported on the canvas and is forced off.

Mistral: Classify Chat

Run whole conversations through one of your custom classifier models: classify a chat transcript by intent, topic, resolution status, or any taxonomy the classifier was trained on, using the full multi-turn context.

Mistral: Classify Text

Run raw text through one of YOUR custom classifier models (trained with Mistral's classifier factory fine-tuning): intent detection, sentiment, routing labels, or any custom taxonomy. Returns per-label score maps per input at data.results[i].

Mistral: Create Batch Job

Start an asynchronous batch inference job at roughly half the price of live calls. Feed it either uploaded .jsonl input_files or - canvas-friendly - inline requests (a JSON array of {custom_id, body} objects), pick the target endpoint and model, then poll with Get Batch Job and download results with Download File Content.

Mistral: Create Embeddings

Turn text into embedding vectors with mistral-embed or codestral-embed (for code). Send one string or a JSON array of strings, get one vector per input at data.data[i].embedding plus token usage. Keep batches modest: every vector is a large float array in the execution record.

Mistral: Delete File

Permanently delete a stored file from your Mistral workspace. This cannot be undone; download anything you still need first.

Mistral: Delete Model

Permanently delete a fine-tuned model. This cannot be undone: prefer the Archive node when the model might be needed again.

Mistral: FIM Code Completion

Fill-in-the-middle code completion with Codestral: give the code before the cursor as prompt and optionally the code after it as suffix, and the model generates what goes between. With no suffix it completes the prompt forward. Output lands at data.choices[0].message.content. Streaming is forced off.

Mistral: Moderate Chat

Moderate a whole conversation instead of raw text: the model classifies the LAST message in each conversation against the nine moderation categories using the earlier turns as context, which catches conversational harms plain text moderation misses.

Mistral: Moderate Text

Classify raw text against Mistral's nine moderation categories (sexual, hate and discrimination, violence and threats, dangerous and criminal content, self-harm, health, financial, law, PII). Returns per-category booleans and scores per input at data.results[i], ready for a Filter node to gate on.

Mistral: OCR a Document

Run Mistral OCR on a PDF or image by URL: each page comes back as clean markdown at data.pages[i].markdown with optional extracted images, tables, headers/footers, and structured annotations via JSON schema. The document-understanding workhorse: point it at an invoice, contract, or scan and pipe the markdown into a chat node.

Mistral: Unarchive Fine-Tuned Model

Restore an archived fine-tuned model to active use. The reverse of the Archive node (the API models unarchive as a DELETE of the archive marker).

Mistral: Update Fine-Tuned Model

Rename a fine-tuned model or update its description. Only applies to your ft:... models, not base models.

Connecting Mistral takes one credential

Mistral authenticates with a single API key created in the Mistral AI Studio console at console.mistral.ai under API Keys. Keys belong to a workspace, and the workspace's plan and rate limits govern every key in it. There is no second field and no OAuth flow to complete.

1

In the Mistral AI Studio console at console.mistral.ai, open API Keys, create a key, and copy it when it is shown.

2

In Mattering Automations, open Connectors, choose Mistral AI, and create a credential holding the key.

3

Click Test Connection. The connector calls Mistral's models endpoint, which verifies the key without spending a token.

4

Click Connect. All 24 Mistral nodes appear in the canvas palette.

Common uses for the Mistral MCP

Generation steps that keep a workflow moving

Drafting, rewriting, summarizing, and structured extraction as one node between your data source and its destination. The chat completion node supports Magistral reasoning models, image input on vision models, JSON mode and structured outputs, and function calling, and Codestral handles code completion through its own fill-in-the-middle node.

Documents in, clean markdown out

Point Mistral: OCR a Document at a PDF or image URL and each page comes back as markdown, with optional tables, extracted images, and schema-driven annotations. Pipe the markdown into a chat node and an invoice, contract, or scan becomes structured fields a workflow can act on.

Bulk work through batch jobs

Queue hundreds of requests as one asynchronous job, which Mistral's docs price at roughly half the cost of live calls. The create node accepts inline requests so no file upload is needed, the get node reports progress and can return outputs inline, and the download node hands back results with each JSONL line already parsed for a For Each loop.

Moderation gates on user content

Classify text or whole conversations against Mistral's nine moderation categories and get per-category booleans and scores, ready for a Filter node. The chat variant judges the last message with the earlier turns as context, which catches conversational harms plain text moderation misses.

Routing with your own classifiers

If you have trained a classifier on Mistral's platform, the classify nodes run raw text or full conversations through it: intent detection, sentiment, resolution status, or any taxonomy you trained, feeding branch conditions in the same workflow.

Embeddings for search and matching

Turn text into vectors with mistral-embed, or code with codestral-embed, one node call per batch of passages. Downstream steps can score similarity for deduplication, retrieval, or matching without leaving the canvas.

Example: An inbound document becomes structured data and a task

Built on the Automations canvas in a few minutes, or described in plain language to Matter and built for you.

1

Trigger: a scheduled run or webhook delivers the URL of a newly received PDF, such as an invoice hitting a shared inbox.

2

Mistral: OCR a Document reads the PDF and returns each page as clean markdown.

3

Mistral: Chat Completion runs in JSON mode over that markdown and extracts the vendor, total, and due date as a structured object.

4

Mattering Productivity: a task is created carrying the extracted fields, so the document ends as tracked work instead of an unread attachment.

Mistral MCP questions, answered

Can an AI agent run Mistral models through this MCP?

Yes. Matter and any MCP-connected assistant, including Claude and ChatGPT, can generate with the chat and Codestral nodes, run Mistral OCR over a document, create embeddings, moderate or classify content, and submit, poll, and collect batch jobs. They use the same 24 nodes the automation canvas uses, under the same credential.

Does this work with Claude, ChatGPT, or other AI assistants?

Yes. Mattering exposes these nodes over MCP, an open protocol for connecting AI assistants to tools, so any MCP-capable assistant can connect to your workspace and reach Mistral through them. Matter, the assistant built into the platform, uses the same nodes natively. Nothing here is tied to one assistant: the point of the connector is to give whichever assistant you already work with access to Mistral's models.

How does this relate to Mistral's own MCP support?

Mistral has embraced MCP from the model side: its agent tooling can call out to MCP servers as tools. This connector runs the other direction. It makes Mistral's API itself available as MCP tools, so the assistant you already use can call Mistral alongside every other connector in your workspace, with the credential stored once and every call visible on a canvas you can audit and edit.

Is there a free way to test it?

Yes, with a caveat you should weigh. Mistral's free plan (Free mode) lets you create an API key with no credit card required, with included monthly usage suited to evaluation and prototyping. Mistral states it may use free-plan API data, input and output, to train its models unless you opt out in the Admin Console, so keep free-tier calls to synthetic or test data and move real workloads to the pay per use Scale plan, which is opted out of training by default and lifts rate limits as well.

What are the rate limits?

Mistral enforces limits per workspace, shared by every key in that workspace, on two axes: requests per second and tokens per minute and per month. The exact numbers depend on your plan and tier and are shown in Mistral's admin console rather than published, so this page will not quote figures. When a call is throttled, the node returns the 429 error body with success set to false, so a workflow can branch on it or retry on a later scheduled run.

Which models can the nodes call?

Whatever your key's workspace has access to. Model fields are free text with documented -latest defaults rather than fixed dropdowns, because Mistral ships new generations and retires old ones on a published deprecation schedule. Run Mistral: List Models for the current roster, with each model's capabilities, context length, aliases, and deprecation date, or Mistral: Get Model for one model's details. Your fine-tuned ft: models appear there too.

What is not covered by the 24 nodes?

A few gaps are deliberate. File upload needs a real binary multipart payload the canvas config panel cannot produce; batch jobs still run end to end because the create node accepts inline requests, and everything else in the files API is shipped. Mistral's legacy fine-tuning jobs API is listed as deprecated in its docs, so no nodes were built on it, though managing existing fine-tuned models is covered. The newer Agents and Conversations APIs are marked beta and are skipped for that reason, but the stable agent completions endpoint is shipped, so an agent built in Mistral's console can be run by ID today. Audio endpoints and streaming are out of scope: nodes return one complete response.

What credential does the connector use, and who bills the usage?

One Mistral API key, sent as a bearer token and stored encrypted in Google Secret Manager. Mistral bills token, OCR, and batch usage to that key's workspace on your own Mistral account, under the plan and limits you hold there. That is separate from Mattering credits, which are pooled at the organization level and cover running the automation itself.

24 nodes waiting on the canvas

Put Mistral to work inside your workspace

Connect once, then use every node in visual automations and AI agent workflows across the Mattering platform.

All product names, logos, and brands are property of their respective owners and are used for identification purposes only. Use of these names, logos, and brands does not imply endorsement or affiliation. Mistral is a product of its respective owner; Mattering is an independent platform.

Start Free Explore our Products