Privacy Policy

ALKARTIS is operated by ALKARTIS LLC (“ALKARTIS,” “we,” “us”), which is the entity responsible for the information practices described in this policy.

ALKARTIS is an intelligence platform that aggregates publicly available government records — property, licensing, employment, and civic data published by the City of Philadelphia and other public sources — and organizes them into readable profiles, timelines, and reports for our customers. We do not create, author, or independently verify the underlying official records; we collect, structure, and present data that governments have already made public.

Public government records. Depending on the app, this includes property assessments and sales, code violations, permits, business licenses, rental licenses, tax delinquency status, recorded deeds, and quarterly public-employee salary and title records. We collect these because they are the underlying subject of our products — an app like LandlordEye or CompensationAtlas has nothing to show without them.

Account data. Your name, email, organization, and password (stored as a salted hash, never in plain text) so you can sign in and so we can apply the correct entitlements to your account.

Usage data. Sign-in history, feature usage events (e.g. creating a watch, generating a report), and coarse activity timestamps, so we can operate the product, enforce plan limits, and diagnose problems.

Review request data. If you submit a review request from a profile page’s “Data Sources” footer, we collect the request type, your description, and your account’s contact email so our team can investigate and follow up.

To understand which parts of the platform are useful and to decide what to build and improve, we record lightweight, first-party usage analytics while you are signed in: the app and feature you are viewing and how long you spend in each. This is tied to your account.

What we deliberately do not collect. We normalize every path before storing it, so we keep only the app and a general feature name (for example, “CompensationAtlas → employee profile”) — never the specific record you looked at, your search terms, your IP address, or your device details. The analytics log is not a browsing history and cannot be used to reconstruct one.

First-party and access-controlled. This data stays in our own systems — we do not send it to any third-party analytics provider, and we never sell it or use it for advertising. Only ALKARTIS administrators can view it, and any view of an individual account’s usage is itself logged for review.

Your choices. You can turn usage analytics off for your account, and download or permanently delete the usage data tied to your account, at any time from your account settings. Organization and enterprise administrators can disable it for an entire organization. Raw usage events are automatically deleted after a limited retention window (180 days by default).

The Data Center (alkartis.com/data-center) publishes a weekly summary of what changed in each city’s public records — totals for the city and its areas, compared with what is normal for the time of year. Its pages are public and can be read without an account.

What it publishes. Every sentence is filled in from public records by a fixed template. It reports totals for areas, not stories about individual people: it never names a private individual, and anything that names a business or a single property is checked by a person before it is published. Corrections and removal requests go through the same review process described below.

The weekly email. You can subscribe without an account. We keep your email address, the city you chose, the area you chose (optional — a ZIP code or district, used only to put that area first), and your language. If you type an address to find your area, we use it once to look up the area and do not keep it. We use this only to send the email; we do not sell it, share it, or use it for advertising. Every email has an unsubscribe link, and unsubscribing deletes your subscription. If you have an account, the email follows your notification settings instead, and “your area” comes from the places you already watch.

Counting visits to public pages. For people who are not signed in, we count visits to public pages in aggregate — which page, which site linked to it, and the country the visit came from — so we know what is read and where readers come from. We do this without cookies, without an identifier that follows you from visit to visit, without storing your IP address, and without third-party analytics or advertising trackers. Signed-in usage is covered by “Usage analytics” above.

If you use the ALKARTIS Assistant, we save your conversations — the questions you ask, the Assistant’s replies, any rating you give a reply, and technical details of how each reply was produced (which ALKARTIS lookups it ran, and any draft reply our accuracy checks stopped). This is different from the usage analytics above, which never record what you type: we keep Assistant conversations specifically to find where it gets things wrong, refuses questions it should answer, or misses data people ask for, and to improve it.

Who can see them. You can see your own saved conversations. ALKARTIS administrators can review them to improve the Assistant, and every time an administrator opens someone’s conversation it is logged. Conversations may be exported, without your name or email, to build test sets for the Assistant. We never sell them or use them for advertising.

Language-model providers. To answer, your question and the ALKARTIS records needed to answer it are sent to the language-model provider configured for your account or organization (for example Anthropic, OpenAI, Google or xAI), under that provider’s API terms.

Your choices and retention. You can delete all your saved Assistant conversations at any time from Account → Assistant. Otherwise they are deleted automatically after a retention window (365 days by default). Please don’t put personal information you don’t want kept into the Assistant.

Some of what you see is derived, not published verbatim by the source. For example, LandlordEye groups properties under a common owner using a name/address matching process it labels with a confidence level (high, probable, or low), and CompensationAtlas turns a person’s quarterly payroll rows into a career timeline of hires, promotions, transfers, and separations using rule-based classification. These derivations can be wrong — a name match can conflate two different people, or a data gap can produce a misleading pattern. That is exactly what the review-request process below exists to catch.

Every profile page generated from public records includes a “Data Sources” footer with a link to submit a review request. You can use it to report:

• A correction request — something shown is factually wrong.
• A match dispute — a profile has been matched to the wrong person or entity.
• A duplicate report — the same subject appears as more than one profile.
• A privacy request — a question about how information is used, or a request under applicable privacy law.
• An identity verification — to support a correction or privacy request made about yourself.

Every request enters an internal review queue and every decision made on it — approved, denied, or otherwise — is logged with a timestamp, the reviewer, and their reasoning. That log is not discarded when a request is resolved; it is the audit trail we use to keep this process accountable.

We do not offer blanket removal of public records on request. The records underlying ALKARTIS are lawfully public, and removing a real record because someone would prefer it not be found would make the platform less accurate for everyone else relying on it — that runs against the trust we’re trying to build.

We will remove or suppress information from a profile when:

• A verified data error exists — the record was mismatched to the wrong person or entity, duplicated, or otherwise doesn’t actually reflect what the source data says.
• We are subject to a legal requirement to do so (e.g. a court order, a statutory sealing or expungement that applies to the underlying record, or a formal request from the government source that published it).
• The underlying government source itself corrects, retracts, or removes the record, and our next scheduled refresh has not yet reflected that.

Outside of these circumstances, our default is to correct or add context to a profile rather than delete it — for example, noting a disputed match rather than silently dropping the record it was based on.

Account data is retained for as long as your account is active, plus a reasonable period for legal and operational purposes after closure. Passwords are hashed, never stored or logged in plain text. Ingested public records are refreshed on a schedule specific to each source (shown on each app’s Data Sources page) and are not retroactively altered except through the correction process described above.

Questions about this policy, or about a specific profile, can be raised through the review-request flow on the relevant profile page, or by contacting your account administrator. See also our Terms of Service.