What is MistralAI-Training?
It crawls web content to build datasets for training Mistral's generative AI models. Mistral says it is not used for search indexing or live answers.
Collects pages to train language models. Blocking it keeps your pages out of future training data, and does not remove you from AI search answers.
Mistral publishes IP lists for MistralAI-User and MistralAI-Index but not for MistralAI-Training.
MistralAI-Training user agent
Mistral AI documents this user agent for MistralAI-Training. Version numbers in it change over time, so match on the name, not the whole string.
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; MistralAI-Training/1.0; +https://docs.mistral.ai/robots)Should you block MistralAI-Training?
Block it if you don’t want your pages used to train models; nothing else changes. Allow it if you’d rather future models know your product, docs or writing.
How to block MistralAI-Training in robots.txt
Add a group for its token to robots.txt at the root of your site.
A bot follows only the group that names it, so it ignores your User-agent: * rules once it has a group of its own. Repeat any paths you also want kept from it there.
User-agent: MistralAI-Training
Disallow: /How to verify a request is really MistralAI-Training
Anyone can send any user agent, so the string alone proves nothing.
Mistral AI publishes no IP ranges or DNS names for MistralAI-Training that we could find, so there is no reliable way to tell a real MistralAI-Training from a scraper using its user agent.
See when MistralAI-Training reads your site
Bots don’t run JavaScript analytics, so a script tag never sees them. invisit records crawler reads on the server instead, through its server SDK or a Cloudflare Worker, and shows which bots read which pages, next to your visitors.
Checked against Mistral AI’s own documentation on . Something changed? Write to help@invisit.dev.