Mistral Vibe

Mistral Vibe is Mistral AI's unified agent platform, created by rebranding and re-architecting Le Chat in May 2026. It gives enterprise teams one agent runtime across two modes — Vibe for Work, which connects to Google Workspace, Outlook, SharePoint, Slack, GitHub and Jira, and Vibe for Code, a coding agent in the terminal, IDE and browser. It is aimed at European and regulated buyers who need sovereignty, on-premises deployment and data residency rather than frontier benchmark leadership.

Mistral Vibe is the French lab's unified agent platform, created by rebranding and re-architecting Le Chat — the conversational assistant Mistral launched in February 2024 — and announced at the company's AI NOW Summit in Paris on 28 May 2026. It splits into two modes running on one shared agent runtime, so connections, tools and user context carry across both. Vibe for Work is a web and mobile agent that connects to Google Workspace, Outlook, SharePoint, Slack, GitHub, Jira and more than 100 other tools, plus any MCP server, to summarise email, analyse spreadsheets and databases, generate PDFs, decks and reports, prepare meetings with attendee context, and execute multi-step workflows after user approval. Vibe for Code is a coding agent available through the web, a VS Code extension, JetBrains and Zed integrations, and a CLI, capable of scaffolding features with tests and documentation, fixing bugs, refactoring, shipping pull requests and running async overnight. The lineage predates the rebrand: Mistral shipped an open-source Vibe CLI alongside Devstral 2 in December 2025, then Vibe 2.0 in January 2026, and remote agents with Mistral Medium 3.5 in April 2026. Alongside the platform Mistral announced Mistral for Industrial Engineering — language models coupled to physics simulation acquired with Emmi AI in May 2026 — naming Airbus, BMW Group and ASML as customers, and Mistral Compute, a €4 billion datacentre programme targeting 200 MW by 2027. Enterprises can deploy Vibe on-premises, in a private cloud, or on Mistral Cloud with full data residency, and fine-tune through Mistral Forge. Mistral holds SOC 2 Type II and ISO 27001/27701 certification.

Category

AI Agents & Orchestration

Pricing

Freemium, Subscription, Contact for pricing

Target Market

CIOs, CTOs, Enterprise Developers, Heads of AI, Platform Engineering Leads

Deployment

Cloud-first, Self-hosted, Hybrid, API-based

Founded

2023

Headquarters

Paris, France

Team Size

500+

Customers

Named enterprise customers include Airbus, BMW Group and ASML; total customer count not disclosed

• ✓
  Vibe for Work

  Web and mobile agent that summarises email, analyses spreadsheets and databases, drafts reports and decks, and executes multi-step workflows after user approval

• ✓
  Vibe for Code

  Coding agent in the CLI, VS Code, JetBrains, Zed and the browser that scaffolds features with tests, refactors and opens pull requests, including async background runs

• ✓
  100+ connectors plus full MCP support

  Native integrations with Google Workspace, Outlook, SharePoint, Slack, GitHub and Jira, plus any Model Context Protocol server, so one agent reaches the existing SaaS estate

• ✓
  On-premises and private-cloud deployment

  Enterprise customers can run Vibe in their own datacentre, a private cloud, or Mistral Cloud with full data residency — the core differentiator versus US-hosted rivals

• ✓
  Shared agent runtime across surfaces

  Chat, CLI, IDE and mobile share the same connections, tools and user context, so switching surface does not mean rebuilding integrations or re-establishing context

• ✓
  Mistral Forge fine-tuning

  Enterprise tier supports custom models, agents and workflows fine-tuned on proprietary data, alongside audit logs, SAML SSO and white-labelling

• ✓
  Mistral for Industrial Engineering

  Couples language models with physics simulation from the Emmi AI acquisition for aerospace, automotive and semiconductor engineering workflows, with Airbus, BMW Group and ASML named as customers

• •
  Sovereign enterprise assistant

  A European bank or defence supplier deploys Vibe on-premises so staff get an agentic assistant over email, SharePoint and Slack without data leaving national jurisdiction

• •
  Background software delivery

  Engineering teams hand routine bug fixes and refactors to async Vibe agents that run overnight and open pull requests for human review the next morning

• •
  Field service diagnostics

  ASML applies Mistral's stack to field service diagnostics, with Mistral claiming solutions arrived 120 times faster than the prior manual process

• •
  Crash simulation reasoning

  BMW Group uses multimodal reasoning over crash simulation data, combining language models with physics rather than fine-tuning a generic model on engineering documents

• •
  Cross-tool research and reporting

  Analysts ask one agent to gather evidence across web sources, internal SharePoint and spreadsheets, then produce a formatted brief or deck in a single pass

List pricing is published for individuals and small teams: Pro is $14.99/month ($5.99 for verified students) and Team is $24.99 per user/month with a $50/month floor, both bundling monthly API credits ($15/mo on Pro). Enterprise is quote-only and is where the things a CIO actually needs sit — custom models and agents, audit logs, SAML SSO, white-labelling, and private or on-premises deployment. Self-hosting and data residency are therefore an Enterprise-tier commercial conversation, not a self-serve toggle, and Mistral publishes no per-seat or per-node figure for it.

This page was written from 7 sources, 5 on domains other than mistral.ai.

1. 1.mistral.ai — vibevendor
2. 2.mistral.ai — pricingvendor
3. 3.help.mistral.ai — 347638 do you have soc 2 or iso 27001 certification
4. 4.venturebeat.com — mistral ai launches vibe expands into industrial ai and anno
5. 5.futurumgroup.com — mistral ai shifts to full stack strategy with vibe and indus
6. 6.hn.algolia.com — hn.algolia.com
7. 7.en.wikipedia.org — Mistral AI