Fioria
Privacy-first educator AI
AI for teachers. Privacy for students.

A teacher-facing AI assistant that never ships student identities to the model.

Fioria gives educators a normal chat experience while adding a privacy transparency layer that shows exactly what the LLM saw and exactly what it returned.

Normal chat experience for teachers
Transparency panel showing exact scrubbed I/O
Browser-memory-only placeholder mapping
Built for FERPA-sensitive educator workflows
Teacher sees
Sarah Martinez is struggling with reading comprehension and needs a new IEP goal for next quarter.
Here is a draft IEP goal for Sarah Martinez with measurable progress language and accommodations.
Transparency window
Sent to model
[STUDENT_1] is struggling with reading comprehension and needs a new IEP goal for next quarter.
Raw model response
Here is a draft IEP goal for [STUDENT_1] with measurable progress language and accommodations.
How it works

A clean, four-step privacy pipeline teachers can actually trust.

1

Write naturally

Paste IEP notes, parent emails, intervention plans, or lesson ideas without rewriting them first.

2

Scrub before send

Fioria detects names, emails, phone numbers, student IDs, grades, and other sensitive education data directly in the browser.

3

AI sees placeholders only

The model receives scrubbed text like [STUDENT_1] and [EMAIL_1]. The mapping never leaves browser memory.

4

Restore locally

Responses come back with placeholders intact, then Fioria restores the real values on-device before display.

Privacy proof

The right panel is not a debug console. It’s the trust layer.

District IT directors can inspect exact scrubbed payloads, while teachers keep a normal conversation view with real names restored locally.

Original text

Jordan Lee (ID 445991) receives speech services twice weekly. Parent email is jordan.parent@example.org.

Scrubbed payload

[STUDENT_1] (ID [STUDENT_ID_1]) receives speech services twice weekly. Parent email is [EMAIL_1].