Privacy Architecture

The AI detector that cannot access your students' work

Impronta · The Morelli Index · Updated September 2026

What happens to submitted text

When you submit a document to Impronta, what happens to it?
Text enters our server over HTTPS. It is processed in RAM only — never written to disk, never logged, never passed to a database. Twenty-six features are extracted. A score is computed and returned to you. The text ceases to exist in our systems the moment the response is transmitted. Not eventually. Immediately. There is no deletion schedule because there is nothing to delete.

Step 1

Text enters server

HTTPS only · encrypted in transit

Step 2

RAM only

Never written to disk · never logged

Step 3

26 features extracted

Metadata only · text never leaves RAM

Step 4

Score returned to you

Feature values + interpretation

Result

Text ceases to exist

Not scheduled for deletion — never stored

"Impronta's zero-retention design is not a policy choice. It is an architectural consequence. You cannot delete data that was never stored. Submitted text is never logged, never backed up, never retained in any form. No Impronta employee can access it — because it is not there."

What we retain (per account) vs. what we never retain

Item	Impronta	Turnitin	GPTZero	Originality.ai
Submitted student text	✗ Never stored	✓ Stored in database	~ Stated not to train on	✓ May be used for training
Analysis score (metadata)	✓ Per account	✓ Per account	✓ Per account	✓ Per account
Feature values (26 metrics)	✓ Per account	Not applicable	Not applicable	Not applicable
Used to train AI models	✗ Never	✓ By policy	~ Stated opt-out	✓ By default
FERPA compliant	✓ Architecturally	~ Via BAA	~ Claimed	~ Unverified
GDPR compliant	✓ No data to delete	~ Via DPA	~ Claimed	~ Unverified

Technical architecture

Impronta's analysis pipeline runs as a stateless serverless function hosted on Netlify's CDN edge network. Serverless functions are ephemeral by design: the execution environment is created for the duration of a single request and destroyed immediately after the response is sent. There is no persistent file system, no writable database connection, and no mechanism by which text could be stored even if the code attempted to do so.

The 26 CNI features extracted from submitted text are returned to the user's browser and optionally stored in the user's own browser localStorage — under the user's control, on the user's device. Impronta's servers never see these feature values after the response is transmitted.

What "zero retention" means for instructors

When a student's essay is submitted through Impronta, no copy of that essay exists anywhere in Impronta's infrastructure after the score is returned. An instructor cannot be compelled, through a data subject access request or legal discovery process, to produce student writing submitted to Impronta — because Impronta does not have it.

This is architecturally different from competing systems that retain submitted text in databases used to detect future plagiarism or train their detection models. Impronta does neither.

Sources: Turnitin Privacy Policy (Feb 2026) · GPTZero privacy policy · Originality.ai terms · Competitor claims marked "Unverified" indicate a privacy claim was made but is not independently auditable.
This document reflects the architecture of Impronta v5 as deployed on Netlify serverless infrastructure. The described properties hold as long as the current serverless deployment model is maintained.