Grammarly
Grammarly
Careful
Grammarly's trust page says you can stop it from using your content to train models by turning off Product Improvement and Training. The privacy-policy URL now opens Superhuman's privacy policy, which says you can turn off the use of user content to train AI models in account settings. That is an opt-out, so the browser extension is Careful on work pages until the control is off. The trust page does not say the control ships off.
Last checked 2026-11-10
Data-safety checklist
Trains on your data by default
Unknown
Both pages describe an opt-out of training on your content. Neither page we read says the control is off by default.How to opt out
Steps
Turn off Product Improvement and Training in Grammarly account settings. The trust page also describes a separate tailored-assistance opt-out at account.grammarly.com/customize/features. Documents saved in the Grammarly editor stay until you delete them or the account.Data retention
What they keep
Saved editor documents are stored until you delete the document or the account. A retention period for extension text that is not saved as a document is not stated.BAA / HIPAA
Unknown
The trust and privacy pages we read do not mention a business associate agreement.Certifications
1 listed
- SOC 2 and ISO certifications are referenced on the trust page. The page does not name the ISO numbers in the text we read.
Enterprise and admin controls
Controls
The trust page says more than 50,000 organizations use Grammarly. It does not, in the section we read, document an admin-wide training lock. The training control is described as an account setting.Sources
2 pages