Mistral AI to add Dictation, Memory, Projects and Research tools to Le Chat

Mistral AI’s upcoming feature set, as revealed by hidden flags and code references, shows the company moving rapidly to bring its Le Chat platform in line with top competitors. The freshly announced Voxtral audio models will soon enable users to dictate messages directly in the chat, transcribing speech to text with an animated interface styled in the brand’s signature look. While public access is not live, the feature’s integration hints at an imminent launch across web and mobile.

This development targets users who prefer or require voice input, aligning with accessibility and hands-free trends seen on platforms like ChatGPT and Gemini. The transcription feature is expected to function similarly, with an emphasis on visual appeal thanks to its themed animation.

Alongside voice input, Mistral is experimenting with a new Research feature. Two modes, "DEEP_RESEARCH_COMPLETION_MODEL_NAME" and "FAST_DEEP_RESEARCH_COMPLETION_MODEL_NAME", are referenced in the code, but their specifics remain undisclosed. The interface will soon sport a Research button, designed to generate research plans after asking the user for clarifying details, drawing inspiration from flows pioneered by Gemini and OpenAI. While non-functional for now, this signals a future where Le Chat can offer structured research assistance.

Memory management is also in early stages, currently marked as Alpha in the sidebar. Users will have granular control, able to enable or disable the feature per chat, and manage stored memories for better personalization or privacy. Organizational tools are expanding as well, with Projects and custom prompts in the pipeline, echoing established solutions from ChatGPT.

These initiatives fit Mistral’s strategy of rapidly closing the feature gap with larger AI players while building distinct user-facing options. As these rollouts unfold, Mistral positions itself as a more complete AI assistant, catering to a wide user base from casual chatters to power users needing structured research and memory control. No firm timelines are set for most features except voice input, which is expected soon. The progression reflects Mistral’s broader move to compete directly with market leaders through consistent feature upgrades.