Vibe by Mistral (ex-Le Chat)

Vibe splits into two modes that share one product. Vibe for work handles research, documents, and automation, while Vibe for code writes, tests, and deploys across your codebase. You pick the mode that matches the task, and your conversations and settings stay in one place.

Vibe reads from the tools your team already uses, including email, calendar, Slack, GitHub, Jira, Notion, and Linear. Teams running their own services can add any remote MCP server, so context flows in without manual copying. This is what turns a chat assistant into something that can act on real data.

Coding agents run in isolated sandboxes for parallel tasks, so several jobs move at once. The teleport feature hands a session from your laptop to a remote agent and back, which suits a task you start at a desk and finish later. Run history, tool calls, and code changes stay visible in the observability view.

You can schedule prompts, agents, or workflows to run at set times, and package a workflow into a reusable agent. That covers recurring jobs like weekly briefs or bulk document processing. Free accounts get a small number of scheduled tasks, while paid plans lift the ceiling.

Canvas builds documents, presentations, mockups, and code inside Vibe. You can upload spreadsheets or connect databases, spot patterns, and turn findings into a deliverable without leaving the app. Supported uploads span PDF, images, CSV, text files, and Office formats.

Voice mode records or uploads audio and returns transcriptions, summaries, and answers. It runs on Voxtral, Mistral's own speech model, which keeps the feature inside the same product rather than a separate app. This helps when you want to talk through an idea faster than you can type it.

The Vibe CLI brings a terminal-native coding agent powered by Devstral, and the VS Code and JetBrains plugins add tab completion, natural-language editing, and context-aware chat. Developers who live in the terminal get the same agent without opening a browser tab.

Enterprise customers can deploy Vibe on-premises, in a private cloud, or on Mistral Cloud with full data residency, and add SAML SSO, audit logs, and white labeling. Fine-tuning on your own data runs through Mistral Forge, so the models adapt to your domain while your data stays in your environment.