Search engines have trained everyone to lower the bar.
You type something specific and get a page of SEO‑bait. You refine the query, the engine “corrects” your words, then follows you around the web to sell you things based on what you just searched for.
Somewhere along the line, “search” stopped being about understanding your question and started being about understanding you—in the behavioral, profile‑building sense.
Momor was built as a reaction to that.
The Line We Draw
Momor is built to understand what you’re asking so it can decide what to run, pull the right sources, and return something useful.
If you search for:
- “can I mow my lawn today”
- “is it safe to drive to Denver tonight”
- “how much vitamin D is too much”
we have to read those words, match intent, and pull the right sources. There’s no honest way to do that without seeing the query.
But there is a line we refuse to cross:
- We don’t turn those queries into a behavioral profile of you.
- We don’t stitch your searches together over time into “this person’s dossier”.
- We don’t use those queries to train AI models.
- We don’t follow you across sites with third‑party trackers.
That’s the core of our Privacy Promise:
We know what was asked, not who asked it.
What We Mean by “Tracking”
“No tracking” gets thrown around a lot. Most of the time, it actually means:
- “We don’t run third‑party ad pixels”
- “We don’t set cross‑site cookies”
Those are good starts, but incomplete. The kind of tracking we care about avoiding looks like this:
- A persistent identifier (cookie, device fingerprint, ad ID…)
- Tied to your behavior over time: searches, clicks, time of day, devices
- Joined with location and third‑party data
- Used to predict what you’ll do next and sell access to that prediction
In other words: tracking as profiling.
We don’t do that. We do something much narrower:
- Short‑lived logs so we can debug “that one weird query” that breaks things
- City‑level location so we can answer “should I bike to work tomorrow”
- Telemetry to know if searches are fast, not who ran them
Those are operational tools, not behavioral dossiers.
How We Actually Handle Your Data
Here’s the concrete picture behind the promise:
Queries and Logs
- Live search data lives in cache for about 1 hour–1 day, depending on the TTL for a given service.
- Application logs that contain query metadata are retained for 30 days, then deleted or irreversibly de‑identified.
- Feedback (thumbs up/down + comments) also lives 30 days, then goes away.
We keep enough history to debug, measure reliability, and improve ranking. We don’t keep a lifetime view of “what you like to search for.”
IP and Location
- We use IP addresses to:
- Derive coarse, city‑level location (e.g., “Naperville, IL”)
- Apply fair rate limiting and prevent abuse
- IPs used in rate limiting are hashed and retained for 24 hours.
- For telemetry, we strip the last octet before anything is stored, so we see “city‑ish” rather than your exact address.
When a query genuinely needs precise location (weather, local timing, routes), we don’t just silently peek at your IP and guess:
- The frontend intercepts the search.
- We ask for location via a prompt with three options:
- Use precise GPS (browser permission, opt‑in, unchecked “remember” box by default)
- Enter a location manually (“Los Angeles, CA”)
- Use IP‑based approximate location
- If you say “Los Angeles, CA,” that’s what we trust.
- GPS coordinates are stored in your browser (session or local storage, depending on your choice), sent only when needed, and never written to our database.
Cookies and Accounts
- Anonymous users get no tracking cookies.
- We use a small number of essential cookies for:
- Authentication (
__Host-cookies) - Keeping you signed in securely
- Authentication (
- No third‑party ad scripts. No cross‑site tracking pixels.
Ads and Paid Plans
Right now, there are no ads. If that ever changes for free tiers, we’re drawing a hard line:
- Free/Basic: may see contextual ads someday (based only on the current page or query, not a profile).
- Paid plans (Pro, Max): no ads, ever.
We’ve wired that into our Terms and product copy so it’s not just a slogan.
Why We’re Writing This Now
If you’ve read our Philosophy and early updates, you’ve seen versions of this idea before:
- “We see queries, not identities.”
- “We never train our AI models on your search queries.”
- “No cross‑site tracking or third‑party tracking cookies for anonymous users.”
The behavior in the code already matched those lines, but the wording across the site wasn’t always as precise as it should be. You’d see things like “No tracking, ever” in one place and a detailed retention table in another.
So we:
- Tightened the language in the Privacy Policy and Terms of Service
- Updated pricing, About, Roadmap, and Settings → Privacy to use the same vocabulary
- Wrote and published a formal Privacy Promise section you can link to
This post exists because we didn’t want to quietly edit old content and pretend it always said what it says now. Instead, we’re telling you exactly what changed and why.
What This Means for You as a User
Day to day, here’s the experience we’re aiming for:
- You type what you mean; we don’t cross out your words or second‑guess your intent.
- We use just enough context (time, place, query) to answer your question, then let that data age out on a predictable schedule.
- We don’t try to predict your personality; we try to understand your question.
When we say:
Search that gets it
We know what was asked, not who asked it
we’re not saying “trust us because we’re nice.” We’re saying: trust us because the system, the policies, and the business model are all aligned with that promise — and if you ever find a gap, we want you to call it out.
If you have questions, concerns, or see something that feels off, email privacy@momor.ai. That inbox exists for exactly this reason.