Le Chat
Free value-added services
Comprehensive List of AI Tools AI dialogue chat

Le Chat

The universal European conversation, research, and enterprise collaboration assistant launched by Mistral AI

Tags:

What is Le Chat?

Le Chat is an AI assistant developed by the French company Mistral AI. In June 2026, the developers officially upgraded Le Chat and gave it a new name.Mistral VibeThe original website entrance, accounts, subscriptions, conversations, and saved content are all retained.

Vibe is not merely a name change for a brand; it represents the integration of everyday conversation tools, professional office agents, and programming agents into a single product.

The current Vibe offers three modes: Work, which is designed for multi-step office tasks across different applications; Code, intended for use with code repositories, terminals, and IDEs; and Chat, which retains the features of the original Le Chat such as turn-based conversations and in-depth discussions. The tool’s directory can continue to use the name “Le Chat” to facilitate searches, but the actual name and SEO settings must indicate that it is Mistral Vibe.

Vibe Work Office Agent

Work is suitable for complex tasks that require multiple steps, various sources, or external tools to be used. After the user describes the goal, Vibe collects the relevant context, breaks down the tasks into steps, invokes the necessary tools, and displays the progress in real time.

Confirmation is requested when performing sensitive actions such as sending emails, modifying data, creating calendars, or deleting content.

Typical uses include summarizing long-term contracts and identifying renewal risks, generating reports from Notion data and customer feedback, researching competitors and creating briefs, organizing unread email drafts, and compiling project progress as well as handling documents. Work offers features such as web search, webpage opening, file uploading, Libraries, Canvas, Projects, Skills, Connectors, and enterprise Workflows.

Connectors and scheduled tasks

The Vibe connector supports tools such as Gmail, Outlook, Slack, Notion, Linear, GitHub, Atlassian, Google Drive, SharePoint, Stripe, and it provides a directory of additional MCP connectors that can be added. Administrators can configure third-party services using OAuth, or they can connect to custom MCP servers as well.

Third-party MCPs are not Mistral’s own products; it is necessary to verify the provider, permissions, and data processing rules before using them.

Work allows tasks to be scheduled to run once, daily, weekly, monthly, or annually. Repeated reports, information monitoring, and routine tasks can be packaged as Skills or scheduled tasks.

Approval and failure notifications should be retained when external writing is involved.

Vibe Code programming Agent

Code is designed for source code, terminals, IDEs, and software engineering tasks; it allows for reading files, modifying code, running commands, analyzing failed tests, and preparing Pull Requests. It offers a local CLI, VS Code extensions, JetBrains plugins, as well as remote Web sessions.

Remote sessions can connect to GitHub repositories, where branches and pull requests are processed in a cloud sandbox for manual review.

CLI supports interactive mode and script/CI mode; it allows reference to project files, limitation of the number of rounds, filtering of available tools, and switching between different approval policies. Vibe displays plans and tool calls, and authorization is required when reading files outside the working directory or performing sensitive operations.

Even with automatic approval, the team should operate in isolated branches and controlled environments.

Vibe Chat and the original Le Chat feature

Chat is suitable for quick Q&A and for users who rely on older features. It retains Agents, Think mode, Deep Research, Python Code Interpreter, and Memories.

Mistral is working on redesigning these capabilities for Work, but there are still differences between the two modes at present.

  • Deep Research:Planning issues, searching multiple sources, and generating structured research reports with citations.
  • Think:It uses inference models to address complex problems, and supports multilingual thinking as well as the use of multiple languages together.
  • Code Interpreter:Python can be run in an isolated environment, where common data analysis packages are used to process files, create charts, and conduct simulations. Work, on the other hand, provides a built-in TypeScript environment.
  • Agents:Create assistants equipped with dedicated commands and tools; in the future, there will be a gradual transition to Skills, sub-Agents, or custom Agents.
  • Memories:It saves long-term preferences and context, and shows which memory has been referenced. Users can close, edit, delete, import, and export them; they can also use invisible conversations that do not read any memories.

Document OCR and file analysis

Vibe allows users to upload PDFs, images, tables, logs, and other digital files; it utilizes Mistral OCR and visual capabilities to extract content, understand complex pages, and answer questions. It is suitable for contracts, papers, financial statements, scanned documents, and technical documentation.

Projects and Libraries can organize files and conversations into a knowledge space representing the project lifecycle.

OCR may still make mistakes with complex tables, handwritten text, low-quality scans, and multi-column layouts. Critical numbers, dates, names, contractual obligations, and references must be checked against the original document.

For corporate confidential documents, it is also necessary to determine the organizational strategy and the location of deployment.

Voice, images, and Canvas

The voice mode is supported by Voxtral’s capabilities, enabling low-latency speech recognition as well as transcription of conversations and meetings. The image function allows for the creation of images, and editing such as removing objects or changing the background can be carried out using natural language, while striving to maintain consistency in the characters and design elements.

Canvas can be used to create and edit documents, display content, or build lightweight applications.

AI-generated images and audio may be distorted; real-person portraits, voices, trademarks, and copyrighted materials require authorization. When distributing such content, it is also necessary to comply with local requirements regarding the labeling of AI-generated material.

Skills, Projects, and Workflows

Projects allow for the storage of specific instructions, files, and conversation contexts related to the same client, team, or topic. Skills consist of instructions and resources, and they are suitable for use with fixed writing styles, checklists, and repetitive processes.

Enterprise developers can also write Workflows in Mistral Studio and publish them to Vibe, so that business users can invoke them from the Work interface.

Workflows are built on a persistent execution engine that supports retrying, tracking, timeout handling, rate limiting, and human intervention. The control plane is managed by Mistral; companies can deploy Workers and data processing tasks in their own cloud, on-premises, or in a hybrid environment, so that the business logic remains within secure boundaries.

Individual and team prices

Package or versionPrices, quotas, and core benefits
FreeIt can be used freely on the web and on mobile devices; access to the latest models, messages, web searches, image generation, and programming sessions is limited.
Pro$
Team$
Education$
EnterpriseCustom quotes are available; it supports custom models, Agents, Workflows, audit logs, SAML SSO, white-labeling, and private deployment.

All plans are subject to fair use restrictions, and taxes and currency options vary by region. Pro is not truly unlimited;

Official examples show that Pro offers significantly better performance in terms of messages, searching, and image generation compared to the free version; the specific levels of performance may vary.

Mistral API Price Overview

Vibe subscriptions and the Mistral Studio API operate under different billing systems. The API is charged based on the usage of input and output tokens, OCR pages, audio, or other endpoints.

As an example given by the authorities, Mistral Large charges $2 per million input tokens and $6 per million output tokens; batch processing typically comes with a 50% discount.

For different models, regions, and service prices, the API pricing page shall prevail.

Developers should check in Studio the costs associated with model calls, OCR processing, fine-tuning, and various endpoints, and they should set team budgets as well as define key permissions. The Vibe Pro subscription, which costs 14.99 dollars, should not be considered an option that allows unlimited API calls.

Open-source and private deployment

Mistral AI offers both open-source weights and commercially hosted models, but the Vibe application itself is not open-source software. Some of the Mistral models can be downloaded and hosted locally, and the licenses for these models vary; some are licensed under Apache 2.0, while others require a separate Mistral license for commercial deployment, creation of derivatives, or use on a large scale.

The enterprise version supports SaaS, VPC, on-premises, or hybrid deployment, and allows for the combination of custom models, connectors, and workflows. This degree of deployability does not mean that all of Vibe’s source code is available publicly; the relevant documentation should indicate that \"the product is not open-source, with Mistral providing some open-weight models as well as options for private deployment.\"

Le Chat (Mistral Vibe) Usage Guide

Complete a basic task.

  1. Clarify the issue, time frame, location, source priority, and output format;
  2. Upload materials for which you have permission to use in Le Chat (Mistral Vibe) or enter search queries;
  3. First, establish a framework using the Vibe Work office agent, and then use connectors and scheduled tasks to add evidence.
  4. It is necessary to distinguish between factual information from the source, the author’s opinions, and AI-generated conclusions.
  5. Check each item for dates, numbers, the original location, and any conflicting evidence;
  6. The conclusions are manually revised, the verification time is recorded, and then they are published;

Create reusable professional workflows

  1. Break down complex topics into four categories of questions: background, data, comparison, and conclusions;
  2. The Vibe Work office agent, connectors, and scheduled tasks, combined with the Vibe Code programming agent, constitute a set of fixed research steps.
  3. Give priority to using the official website, research papers, regulatory documents, and raw data;
  4. A second person is assigned to review conclusions that are considered high-risk;
  5. Save queries, evidence, versions, and unresolved issues;
  6. Re-run after the data changes and update the conclusions;

Which users are it suitable for

  • Researchers and knowledge workers: in-depth research, web searching, long documents and reports.
  • Developers: CLI, IDE, remote sandbox, code modification, and PR process.
  • Data analysts: OCR, execution of Python or TypeScript code, and visualization.
  • Corporate team: Connects email, documents, project tools, and internal processes.
  • Organizations that place emphasis on European suppliers and deployment control.
  • Individual users who need multilingual reasoning, as well as voice and image processing capabilities.

Restrictions and Precautions

  • Vibe will invoke external tools and may modify data; it is necessary to check the permissions and approvals of the connectors.
  • Research reports, OCR results, code modifications, and business recommendations can all contain errors; it is necessary to verify the sources, test the code, and conduct a manual review.
  • Remote programming tasks should be executed in branches and sandboxes; high-level permissions should not be granted directly to the production system.

Frequently Asked Questions

Has Le Chat changed its name?

Yes. In June 2026, Le Chat was officially renamed Mistral Vibe; the original website address, accounts, subscriptions, history, and content were all retained.

What are the differences between the three modes of Vibe?

Work is used for handling multi-step office tasks across different tools, Code is used for managing code repositories and terminals, while Chat facilitates quick conversations, in-depth research, Think, Python, and Memories.

Is Le Chat or Vibe free?

There is a free version available. The Pro version costs 14.99 dollars per month, the Team version costs 24.99 dollars per user per month, and the student plan costs 5.99 dollars per month.

Can Vibe be deployed privately?

The enterprise version supports customized and private deployment options, but the full Vibe product is not open-source software; the specific license terms and deployment requirements need to be verified one by one.

Does the subscription include an API?

It does not include unlimited APIs. The Vibe personal subscription and the Mistral Studio API are charged separately; the fees for these APIs are calculated based on the usage of tokens, OCR services, and endpoints.

©️Copyright notice: Unless otherwise specified, all articles on this site are copyrighted bySharing of AI toolsAll content on this site is original; without permission, no individual, media outlet, website, or organization may reproduce, copy, or otherwise distribute it, nor may they create mirrors of it on servers that are not owned by this site. Otherwise, we reserve the right to take legal action against such parties in accordance with the law.

Tools similar to Le Chat