The Mistral Ecosystem

Update November 2, 2026 — Le Chat becomes Vibe. Since May 28, 2026, Le Chat is now Vibe (still at chat.mistral.ai): a single agent for work and code, carrying over your conversations, settings, and subscriptions. The legacy Le Chat experience remains accessible in Chat mode; in Work mode, agents are now skills (invoked with @), memories are the knowledge base, and deep search is the Deep Research skill. Plans are now called Mistral Free, Pro, Team, and Enterprise. This lesson uses the name Vibe; some menus described here may have changed names or locations.

A Complete AI Product Ecosystem

Mistral AI does more than build language models. The company has built a complete ecosystem of products and services covering all artificial intelligence needs, from simple conversation to complex application development. When you start out, it’s easy to get lost: Vibe, the Console, the API, Forge… which one is for you? This lesson maps out the ecosystem so you know which door to open for every situation.

Vibe — The Conversational Assistant

Vibe (chat.mistral.ai) is Mistral AI’s consumer interface, the equivalent of ChatGPT for OpenAI or Claude for Anthropic. You access it directly from your browser, without any technical configuration—this is where almost all new users start, and it’s the tool we’ll use for most of this course.

What you can do here covers most daily use cases: ask questions and get detailed answers, write and summarize text, translate, analyze images with multimodal models, generate code, perform real-time web searches with Deep Research, or collaborate visually with Canvas. Imagine preparing a briefing document for a meeting: you can upload documents to Vibe, request a structured summary, then refine the wording in Canvas—without ever leaving the interface.

Vibe offers several models to choose from directly in the interface. You can switch between a fast, economical model for simple questions and a more powerful model when your request is complex. We’ll see later how to make this choice intelligently.

The Console — The Control Center for Developers

The Console (console.mistral.ai) is the dashboard for developers and businesses. If Vibe is the storefront, the Console is the back office: this is where you manage your API keys (creation, revocation, access tracking), monitor your token consumption and associated costs, organize your projects by team using workspaces, and configure the models used by your applications. A developer who sees an unusual bill will always start with the Console: it tells them which project consumed what, and with which key.

The API — Programmatic Access

The Mistral API lets you integrate models directly into your applications. Whether you’re building a chatbot for your site, a document analysis tool, or a coding assistant, the API is your entry point: your program sends a request, the model responds, and your application uses the response.

The API exposes several distinct capabilities, each addressing a specific application need. Chat Completions is the core building block: it handles conversations with models, in text, images, or documents—this is what a standard chatbot uses. The Embeddings transform text into numerical vectors, enabling semantic search: finding passages in a document base close to a question, even when they don’t use the same words. Document AI (OCR) extracts content from scanned documents or PDFs—the essential prerequisite for any automated processing of paper archives.

Three capabilities complete the set. Audio covers transcription and text-to-speech, to go from voice to text and back. Function Calling connects models to your external tools: instead of just responding in text, the model can trigger a search in your database or call a third-party service. Moderation, finally, detects inappropriate content and manipulation attempts—a safety net for any application exposed to the public.

Official SDKs are available in Python and TypeScript, and the API remains accessible via cURL for quick integration in any environment.

Forge — The Fine-Tuning Platform

Forge is the platform that lets you fine-tune Mistral models with your own data. Fine-tuning addresses a specific need: getting a model to exhibit a very specific behavior that a simple prompt isn’t enough to guarantee—a particular brand tone, sharp industry vocabulary, domain expertise. A customer service team wanting all its responses to exactly match the company’s communication charter is a typical case. With Forge, you upload your training datasets, launch fine-tuning sessions on Open models, evaluate the fine-tuned model’s performance, then deploy it via the API like any other model.

Vibe CLI — Command-Line Assisted Development

Vibe CLI is a command-line tool that integrates Mistral models directly into the developer’s terminal. Instead of copy-pasting code between their editor and a chat interface, the developer dialogues with the model where they work: they can generate code from a natural language description, refactor and improve existing code, debug an error by pasting the error message directly, or automate repetitive development tasks.

Workflows — Intelligent Automation

Mistral Workflows let you create automated processing chains combining multiple model calls. This is the building block for businesses whose processes require several successive AI processing steps, each building on the result of the previous one.

A concrete example makes the idea clear. A customer request processing workflow could chain like this:

1. Receive a customer email
2. Analyze it to extract intent
3. Search for the answer in a knowledge base
4. Draft a personalized response
5. Submit it for human validation before sending

Each step uses the model differently, and the whole thing runs without manual intervention until final validation.

Ecosystem Overview

Step back, and the ecosystem organizes into three layers based on your profile. The general public uses Vibe, in Free or Pro version. Developers rely on the Console to manage access, the API to integrate models into their applications, Vibe CLI to code with assistance in the terminal, and Forge to create custom models. Enterprises, finally, add Workflows to automate business processes, on-premise deployment of Open models to keep data internal, and dedicated support for technical and commercial guidance.

Key Takeaways

• Vibe is the interface accessible to everyone for conversing with Mistral models
• The Console and API are developers’ tools for integrating Mistral into their applications
• Forge enables fine-tuning to adapt models to specific needs
• Vibe CLI brings AI assistance directly into the developer’s terminal
• Workflows automate business processes by chaining multiple AI calls
• The ecosystem covers all usage levels, from individuals to enterprises