Legal

How ILTY uses anonymous research data

Last updated: April 25, 2026

What we collect, in plain English

Every field that flows to the research collection:

•Pre-session mood (1-10) and post-session mood (1-10)
•Mood change (the difference, e.g. -2 or +4)
•Emotion tags you picked from our 16-option list (e.g. “anxious”, “grateful”)
•Generalized topic tags (NEVER your messages — see below)
•Which companion you spoke with
•Hour of day and day of week, in your local time
•Your timezone offset (the same offset 200M+ people share — never your city)
•App language
•Session length and message count
•A coarse session-number bucket (1, 2-5, 6-10, 11-25, 26-50, 51-100, 100+)
•A random session ID (newly generated for each session — not tied to your account or your local session record)

What we DO NOT collect

✗Your messages (we never see what you actually said)
✗Your name, email, user ID, or device ID
✗Your exact location, city, or precise coordinates
✗Your age, gender, occupation, or any profile detail
✗Your IP address (the document we save doesn't include it; Firebase may briefly process your IP for the network connection but it's never stored with the research data)

How we make this anonymous in practice

1. Random per-document ID

No two sessions can be linked to the same person, even by us with full database access.

2. Closed-vocabulary topic tags

Your conversation topics are filtered through a fixed list of 12 themes (work, relationships, family, health, sleep, anxiety, self-worth, grief, productivity, money, purpose, boundaries) before anything leaves your device. If a topic doesn't match the list, it's dropped — never uploaded.

3. No raw text

Anything not on the field list above stays on your device.

What we publish from this data

We publish occasional research studies at ilty.co/research about patterns across all users — things like “what time of day do AI conversations help most” or “how mood changes vary across companions.” The studies aggregate across thousands of sessions; nothing about any individual is ever identifiable.

Retention

•Raw per-session documents: 90 days, then deleted
•Computed aggregates from the studies: indefinite, because they describe patterns across thousands of users (not individuals)

Why no in-app toggle

This is the same category of anonymous aggregate data Firebase Analytics already collects without a toggle. Adding a per-feature opt-out for genuinely anonymous data implies the other anonymous data is somehow different, when it isn't.

If you're uncomfortable with any anonymous data collection, you can disable it system-wide in Settings → Privacy → Apple Advertising on iOS, which prevents us from receiving Firebase Analytics events at all. (This also disables the research collection on your device, since it shares the same network path.)

Questions or concerns

Reach out via our contact page. We'll respond within 7 days.

If you want us to confirm we don't have data tied to you, the honest answer is: we couldn't find it even if we tried — there's no user identifier attached to anything in this collection. But the contact form is how to reach us anyway.

What we collect, in plain English

Every field that flows to the research collection:

•Pre-session mood (1-10) and post-session mood (1-10)

•Mood change (the difference, e.g. -2 or +4)

•Emotion tags you picked from our 16-option list (e.g. “anxious”, “grateful”)

•Generalized topic tags (NEVER your messages — see below)

•Which companion you spoke with

•Hour of day and day of week, in your local time

•Your timezone offset (the same offset 200M+ people share — never your city)

•App language

•Session length and message count

•A coarse session-number bucket (1, 2-5, 6-10, 11-25, 26-50, 51-100, 100+)

•A random session ID (newly generated for each session — not tied to your account or your local session record)