MistralAI-Training: what this bot is and how to allow or block it
MistralAI-Training is a bot from Mistral AI that collects pages to train AI models. It does not affect Mistral Le Chat answers with search: you can block it without losing visibility in AI search.
MistralAI-Training User-Agent
This is how MistralAI-Training identifies itself in requests to your site. In robots.txt you only write the token — MistralAI-Training.
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; MistralAI-Training/1.0; +https://docs.mistral.ai/robots)
How to allow MistralAI-Training
Add this group to robots.txt at the site root. Close service sections in the same group: a bot with its own group does not read the «User-agent: *» rules.
User-agent: MistralAI-Training Allow: /
robots.txt is not everything. If your site uses Cloudflare or other bot protection, it may serve MistralAI-Training a 403 error or a challenge page even when robots.txt allows everything.
How to block MistralAI-Training
To block the whole site, add the group below to robots.txt. The rule works for bots that respect robots.txt.
User-agent: MistralAI-Training Disallow: /
What happens if you block MistralAI-Training
New model versions will not learn from your pages. Visibility in answers with search does not change — search bots are responsible for that.
How to check whether MistralAI-Training visits your site
Search the web server logs for the token. The User-Agent string is easy to fake, so to be sure, check the IP address against the ranges if Mistral AI publishes them.
grep "MistralAI-Training" /var/log/nginx/access.log
Does your site let MistralAI-Training in
Free 15-second check: we read your robots.txt and request your site as AI bots — you see which ones the server and bot protection let in and which they block.
Other Mistral Le Chat bots
MistralAI-IndexsearchMistralAI-Useron request
Questions and answers
Should I block MistralAI-Training?
It is your call. Blocking MistralAI-Training does not remove your site from AI answers with search, but models will not learn about you in training either. If content is your main asset, block it; if awareness matters, leave it open.
How do I know whether MistralAI-Training visits my site?
Search the web server logs for the token. The User-Agent string is easy to fake, so to be sure, check the IP address against the ranges if Mistral AI publishes them.
Is it enough to allow MistralAI-Training in robots.txt?
robots.txt is not everything. If your site uses Cloudflare or other bot protection, it may serve MistralAI-Training a 403 error or a challenge page even when robots.txt allows everything.