How it works
Per-user, per-account
Your voice is isolated to your account. Never blended with anyone else's.
What's analyzed
Tone, sentence length distribution, vocabulary, signoffs, response patterns.
What's stored vs computed
Pranan stores the message content and metadata it needs, plus the voice profile it derives. Both stay in your account.
The privacy architecture
Pranan analyzes sent mail in three stages. First: extract patterns. Sentence length, vocabulary, signoffs, response timing, tone shifts per recipient. Second: store those patterns as a voice profile in your account. Third: apply the profile when drafting for you, and never share, sell, or repurpose it for any other user or purpose.
To do this, Pranan stores the content and metadata of your synced mail in its database. It is associated with your account, protected by per-user access controls, and deleted when you disconnect Gmail or delete your account. The Privacy Policy lists what is stored and which providers process it.
No AI model is trained or fine-tuned on your email. The profile is a set of patterns passed to the drafting model at draft time. It is yours, full stop.
Try sent mail analysis free
100 drafts a month on the Free plan. Voice profile built from your sent mail.