US Home Address Privacy Report 2026

US home address privacy in public records cannot be evaluated through a single website or a simple question about whether records are public. Property assessments, tax information, recorded land documents and court records serve different administrative purposes. Their systems can offer different combinations of online searches, document images, copies and onsite inspection. A published confidentiality procedure may apply to one of those routes without changing the others. The central task is to identify what an agency documents, which records its guidance covers and what the available evidence leaves unresolved.

This report examines that documentation through three linked datasets. Dataset A reviews state-level address-confidentiality and privacy-program structures across the 50 U.S. states. Dataset B audits a frozen, purposive sample of 50 local jurisdictions, with four mandatory record categories per jurisdiction and 200 system/category rows. Dataset C examines implementation documentation in 20 jurisdictions selected from the frozen Dataset B pool. The datasets address different questions. They are not interchangeable populations, and the local samples are not nationally representative.

The strongest quantitative findings concern documented access conditions among VERIFIED Dataset B rows with known values for the relevant field. Among 102 such rows with known account requirements, 90 did not require an account, or 88.2%. Among 113 VERIFIED rows with known payment coding, 47 required payment in some or all documented contexts, or 41.6%. Those populations differ because the availability of evidence differs by field. Neither result describes all public records, all audited systems or U.S. counties generally.

Access coding also differed between the approved category-specific results. Among 29 VERIFIED property rows with known public-access coding, 24 were coded Yes, or 82.8%. Among 35 VERIFIED court rows with known access coding, 3 were coded Yes, or 8.6%, while 32 were coded Partially. Yes refers to the relevant system within its documented scope. Partially preserves qualifications rather than treating access as either unrestricted or absent. The findings do not establish that property systems offer weaker confidentiality protections than court systems.

Two approved cases illustrate why these distinctions matter. The Harris County Clerk's real-property portal explains that information has been redacted from certain instruments in its online database and distinguishes that database from the official repository. The example shows why a change to online display should not be described automatically as alteration or deletion of the underlying official record. Utah Courts provides the approved example of documentation clarity: its applicable statewide court guidance documented 8 of 8 components under the frozen Privacy Guidance Clarity Score framework. That is a finding about documented instructions, not a measure of protection effectiveness.

The report therefore separates access questions from remedy questions. An account requirement describes an access condition. A payment requirement describes a charge in the coded context. Neither establishes who qualifies for address confidentiality, whether a substitute address will be accepted, whether an existing image can be redacted or whether historical records are covered. Ordinary correction is also a separate question: updating an inaccurate mailing address does not, by itself, establish a route for withholding an accurate residential address.

The analytical controls are deliberately conservative. Substantive headline percentages use VERIFIED rows and known values. Unknown values do not become negative findings, and NEEDS REVIEW rows do not become confirmed results merely because an official source was found. State-program prevalence claims that did not satisfy the analysis requirements are omitted. These restrictions preserve the difference between an interpretable result and a number that would overstate the evidence.

Account requirements

Among 102 VERIFIED audit rows where account requirements could be coded, 90 did not require an account (88.2%). The denominator excludes 34 VERIFIED rows with Unknown account coding and the 64 rows that were not VERIFIED, out of a planned Dataset B population of 200 rows. The finding concerns the coded access route; it does not establish that every service associated with a system is available without registration.

Payment requirements

Among 113 VERIFIED audit rows with known payment coding, 47 required payment in some or all contexts (41.6%). The denominator excludes 23 VERIFIED rows with Unknown payment coding and the 64 rows that were not VERIFIED. The numerator combines Yes and Partially payment coding. It should not be paraphrased as a claim that all searches or every document request require payment.

Court-record access

Among 35 VERIFIED court rows with known public-access coding, 3 were coded Yes (8.6%) and 32 were coded Partially. The denominator is 35 of the 50 planned court slots. These are audit-slot results, not a count of independent court policies. Partially records limitations in the documented access scope and does not imply that all information in a court system is confidential.

Property-record access

Among 29 VERIFIED property rows with known public-access coding, 24 were coded Yes (82.8%). The denominator is 29 of the 50 planned property slots and excludes 2 VERIFIED property rows with Unknown access coding. Yes does not mean that every record image, historical document or protected address is publicly available.

Illustrative cases

Harris County is used only as an illustrative boundary between an online database and the official repository. It is not a prevalence finding and not a statement that every address can be removed.

Utah Courts is the approved documentation-clarity example, documenting 8 of 8 components of the frozen clarity framework. That is a statement about published guidance, not about protection quality.

Why the denominators stay beside the findings

An audit row is a jurisdiction/category slot. A jurisdiction can contribute several rows, and several rows can refer to a shared system. A VERIFIED row may still contain a justified Unknown for a particular field. The account, payment and category-specific results therefore do not describe identical sets of observations. Only four percentages are published: 88.2%, 41.6%, 8.6% and 82.8%. Court Partially counts and the Utah Courts clarity result are reported as counts, because no further percentage was approved for publication.

A residential address can appear in different administrative contexts. A property system may identify a parcel and its assessment information. A tax system may distinguish the property being taxed from a mailing destination. A recorded instrument may contain text and images supplied when a transaction was recorded. Court information may exist in an index, a docket, a filed document or an access service for a particular user group. The study treats these as distinct record environments rather than assuming that a common address field makes their procedures equivalent.

That distinction matters when someone encounters a public-facing search result. The result may display structured information drawn from a database, while a separate image contains the text of an original document. Changing the structured field does not demonstrate that the image changed. Similarly, an index may be available online while access to the associated document requires another step. The report's access findings describe the relevant system and its documented boundaries; they do not describe every possible representation of its contents.

The responsible authority also needs to be identified. A website associated with a county can link to a statewide judicial service or a system run by a different agency. A state-hosted portal does not necessarily establish that the state controls every local confidentiality decision. In some administrative settings, municipal rather than county offices handle the relevant function. The protocol records administrative level and system scope so that a jurisdiction label does not become an unsupported claim about institutional responsibility.

Privacy procedures add another layer. Enrollment in a state address-confidentiality program, acceptance of a substitute address, restriction of online display and redaction of a filed document are different actions. Their eligibility and scope must be established separately. A general explanation of program participation cannot answer every question about a property assessment, tax record, recorded instrument or court filing.

Dataset A is a state-level review covering the 50 U.S. states. It records program structures and supporting sources rather than serving as a count of local agencies. The District of Columbia and territories are outside that dataset's design. All state rows remained NEEDS REVIEW for publication clearance, so Dataset A is used descriptively and no verified-only state prevalence headline is derived from it.

Dataset B contains 50 purposively selected local jurisdictions and 200 mandatory audit rows. Each jurisdiction contributes four categories: property assessor/property records, property tax, recorder/land records and court records. These are planned audit slots, not necessarily separate organizations. The sample was frozen before local policy research so that jurisdictions would not be replaced because their findings were inconvenient, unclear or difficult to research. Dataset B was selected purposively for geographic and intended-size variety. It is not a probability sample and not statistically representative.

Dataset C is a 20-jurisdiction deep dive into state-to-local implementation, drawn from the frozen Dataset B pool using an outcome-blind SHA-256 ordering within frozen region and intended-size quotas. Privacy outcomes, verification status and ease of research were not selection criteria. Only Dataset C used that hash procedure; Dataset B did not.

Across the datasets, the object of study is official policy and system documentation. The research did not submit personal record searches, create access accounts, pay for records or seek to bypass restrictions. Full detail is set out on the methodology page.

The state review distinguishes a main Address Confidentiality Program from additional or narrower statutory programs. That separation prevents an employee-specific shield, a voter-focused procedure and a broader program from being presented as identical coverage. Eligibility belongs to the particular program and its documented scope. A generic label such as "address protection" is not enough to establish that the same groups or records qualify.

A substitute-address model can provide a designated address for use in specified contexts. Acceptance of a substitute address is different from removal of information already held in a record. The protocol therefore treats prospective use, existing-record treatment and the explanation of that treatment as separate questions. A procedure can clearly explain its prospective limits without authorizing retrospective changes.

State-level authority is relevant to a local implementation question, but it does not settle that question. Evidence that a program exists does not establish that a recorder's website links its procedure, that an assessor accepts a particular document or that a court record changes automatically following enrollment. The local procedure must be supported by applicable local guidance or an explicitly applicable shared source.

This report consequently makes no state-program prevalence claim and publishes no program count. All Dataset A rows remained NEEDS REVIEW for publication clearance. The state review's role here is to explain the kinds of policy structures that local implementation research must distinguish, without implying nationwide uniformity or automatic implementation.

The local audit records public access, account requirements and payment requirements as separate fields. A system can permit a public search while requiring an account for another service. A charge can apply to copies rather than to browsing. Onsite inspection and online access can also have different conditions. Keeping those fields separate preserves distinctions that would disappear in a single "open" or "closed" label.

Public-access coding requires the same care. Yes applies to the relevant system in the scope established by evidence. Partially preserves a limitation rather than forcing an uncertain or mixed condition into Yes or No. Neither value should be translated into a claim that a person's address was located. This audit did not perform that test.

Several audited jurisdictions share a system or a policy with other jurisdictions. Shared systems and shared policies were recorded with identifiers and deduplicated in the analysis, so that a single statewide platform used by several counties is not counted as several independent findings.

Property systems were the most consistently documented category in the audit. Among 29 VERIFIED property rows with known public-access coding, 24 were coded Yes (82.8%). Property assessment and parcel systems commonly publish an online search covering ownership and parcel information, and their help pages tend to describe what the search returns.

That consistency should not be read as a privacy conclusion. A high proportion of documented online access means the systems describe an online route, not that a specific residential address is displayed, indexed by search engines, or ineligible for a confidentiality procedure. Two VERIFIED property rows remained Unknown on access coding and are excluded from the denominator rather than recoded.

Harris County: an illustrative boundary case

Harris County illustrates a distinction that runs through the whole report: what a public body shows online is not always the same as what its official repository holds. The Harris County Clerk's real-property portal explains that information has been redacted from certain instruments in its online database, and distinguishes that database from the official repository, which remains subject to separate rules and separate request procedures.

It is presented as an illustrative boundary case, not as a measured rate or a claim about how common the pattern is. It matters practically because a person who sees an address disappear from a portal may reasonably, but wrongly, assume the source record was changed.

Court rows were coded very differently from property rows. Among 35 VERIFIED court rows with known access coding, 3 were coded Yes (8.6%) and 32 were coded Partially. Partially is the dominant coded value because court systems frequently publish an index or docket while restricting document images, remote access or particular case types.

This does not establish that court records are more private than property records. The two results use different eligible row sets and describe documented online access scope, not the sensitivity of the information held or the availability of a confidentiality remedy. A Partially coded court system may still display a party's address in an index entry, and a Yes coded system may withhold protected material.

Court categories also raise the shared-system issue most sharply. Several audited counties use a statewide judicial platform, so the 35 rows are audit slots rather than 35 independent court policies.

Two frequently assumed barriers to public-record access are registration and payment. In the audited rows they behave differently. Among 102 VERIFIED rows with known account coding, 90 did not require an account (88.2%). Among 113 VERIFIED rows with known payment coding, 47 required payment in some or all documented contexts (41.6%).

Those two denominators are not the same set of rows and cannot be combined into a joint access statistic without a separately approved analysis. Payment coding combines Yes and Partially, because a documented fee in at least one context, for example certified copies or bulk downloads, is materially different from a fully free system but is not the same as charging for every search.

For an individual concerned about address exposure, the practical implication is narrow: the absence of a login or a fee in a coded system does not tell you what a system displays, and the presence of a fee is not a privacy protection. Fees and accounts are administrative access conditions, not confidentiality mechanisms.

Access coding answers the question "what does this system document about getting to the record?". A privacy remedy answers a different question: "what can be withheld, redacted, suppressed or substituted, for whom, in which records?". The report keeps these apart deliberately, because access data cannot support remedy conclusions.

Four distinct remedy types appear in the documentation reviewed: correction of inaccurate information, confidentiality or protected-address status for eligible people, redaction or suppression of specified data elements, and restriction of online display only. They have different eligibility rules, different forms and different effects on existing records.

Correction is the remedy most often documented, and it is also the one most easily misread. A route for fixing an incorrect mailing address does not establish a route for withholding an accurate residential address. Where official guidance does not state whether an existing or historical record is covered, the report records that silence rather than inferring either outcome.

Dataset C examined how state-level rules appear, or fail to appear, in local documentation. It covers 20 jurisdictions drawn from the frozen Dataset B pool using the outcome-blind SHA-256 ordering and frozen region and size quotas described in the methodology.

The central observation is structural rather than numerical: a state program can be well documented at state level while the local agency that holds a particular record publishes nothing about how that program affects its own systems. Where that link is missing, the evidence is incomplete. It does not establish that the program has no local effect, and it does not establish that it has one.

Because the deep dive's aggregate scoring results did not clear the publication rules, this report publishes no Dataset C prevalence figure, no local program recognition rate and no aggregate clarity statistic. Only the approved Utah Courts example is published.

The Privacy Guidance Clarity Score is a 0 to 8 documentation measure. One point is awarded for each documented component: a guidance page, a dedicated process or form, stated eligibility criteria, required documents, a submission method, treatment of existing records, treatment of historical records, and contact information for questions.

Utah Courts is the approved example, documenting 8 of 8 components in its applicable statewide court guidance. That is judicial-branch guidance with statewide scope; it is not a county policy, even though the selection identifier for the record references the sampled county.

• It does not show national prevalence. Dataset B is purposive and non-probability, so there is no margin of error and no estimate for U.S. counties generally.
• It does not show that any address is or is not exposed. No person searches were performed and no real residential address was tested.
• It does not show whether a confidentiality request would succeed, how long it takes, or whether an agency applies its own documented procedure consistently.
• It does not show change over time. This edition is a baseline and does not demonstrate that procedures improved or worsened relative to earlier years.
• It does not show that Unknown means No. Unknown means the documentation did not allow a determination and is excluded from denominators rather than recoded.
• It does not show a state-program prevalence figure. Dataset A rows remained NEEDS REVIEW for publication clearance and are used descriptively.
• It does not measure privacy risk, safety, protection quality or effectiveness. The clarity score concerns documentation only.

For individuals, the practical implication is that "is this record public?" is the wrong first question. The useful questions are which agency controls the record, which access route is being described, and whether the documented remedy covers new submissions, existing records, historical material or online display alone.

For agencies, the audit suggests that documentation quality and access design are separable. A system can be easy to search and hard to understand from a privacy perspective, or well documented on confidentiality while offering limited online access. The eight clarity components describe what a complete public explanation looks like without prescribing any particular policy.

For researchers and journalists, the denominators matter more than the headlines. Every figure here is bounded by verification status and known coding, and each of the four published percentages describes a different eligible row set. Re-using them as interchangeable measures of openness would misstate the evidence.

The following framework reflects what the audited systems actually document. It is deliberately neutral, and each step can be carried out by an individual.

1. Identify the record type. Establish whether the address appears in a property assessment record, a property tax record, a recorder or land record, or a court record. Each category is governed by different rules, and requests directed at the wrong category are usually refused.
2. Identify the responsible agency or shared system. Find the body that actually controls the underlying record, which may be a county assessor, recorder, tax office, court clerk, or a shared statewide platform used by several jurisdictions. The website displaying the record is not always its controller.
3. Determine whether the issue is correction or confidentiality. Inaccurate or outdated information may be handled through a correction route. Accurate but sensitive information requires looking instead for a confidentiality, redaction, suppression or protected-address procedure.
4. Check whether an Address Confidentiality Program or other protected-address rule applies. Check the eligibility criteria and whether the program covers the specific record category. Participation in such a program does not automatically change every existing public record, and coverage frequently differs by record type.
5. Distinguish online display from the official underlying record. An online portal, an index entry, a scanned image and the official repository can be separate things with separate procedures. In the Harris County example, a redaction in the online real-property database is not necessarily a change to the official repository.
6. Check existing and historical records separately. Prior filings, archived copies and historical images are often handled under different rules from current records. Where official guidance is silent, do not infer that a change applies retroactively.
7. Treat search-engine de-indexing as a separate layer. De-indexing affects search visibility only. It does not delete or amend the government source record, and the two processes succeed or fail independently.

This is a research-informed framework, not legal advice and not a promise that any request will succeed. Official guidance for your specific jurisdiction and record type remains primary.

Dataset A is a 50-state review of address confidentiality and privacy protection structures. It is structurally complete, but all state rows remained NEEDS REVIEW for publication clearance, so it is used descriptively only. Dataset B is a purposive sample of 50 local jurisdictions with four mandatory categories each, giving 200 audit rows; it is not statistically representative and carries no margin of error. Dataset C is a 20-jurisdiction state-to-local implementation deep dive, selected from the frozen Dataset B pool by outcome-blind SHA-256 ordering within frozen region and size quotas.

Coding rested on official sources: the agency's own page, form, statute or documented procedure. Search-engine snippets were treated as discovery only. Inaccessible evidence was coded Unknown, and Unknown is never converted to No. Shared systems and shared policies were deduplicated. Rows carry one of three statuses, VERIFIED, NEEDS REVIEW or UNVERIFIED, and published percentages use VERIFIED rows with known coding only. No person searches were performed, no accounts were created, no payments were made and no restriction was circumvented.

Full definitions, the frozen protocol, the verification rules and the clarity-score components are set out on the methodology page.

The curated public release contains 201 records: all 200 Dataset B record identities and one approved Dataset C record for Utah Courts. Dataset B status counts are 136 VERIFIED, 53 NEEDS REVIEW and 11 UNVERIFIED. Substantive access, account and payment values are published where the row is VERIFIED; non-VERIFIED rows carry their identity and status only. Literal Unknown values remain Unknown. The four published percentages can be reproduced directly from the file.

No license is claimed or asserted for the dataset, and no DOI has been issued. The HTML edition of this report is the canonical version of record; the PDF is a convenience copy of the same content.

• Curated public release (CSV)
• PDF edition
• Methodology

Suggested citation: Milan Eedel Bloedt, US Home Address Privacy Report 2026. Info Remover, September 10, 2026.

Info Remover is a commercial provider of online content removal, search de-indexing and reputation management services. This report was produced and funded internally and is published as company-funded original research with a disclosed methodology. It is not independent academic research and was not sponsored externally.

The report is informational. It does not provide legal advice, does not assess an individual's circumstances, and nothing in it should be read as a promise that any record, listing or search result can be removed.

1. Published percentages use VERIFIED Dataset B rows with known coding for the field in question. Denominators differ by field because evidence availability differs.
2. Account requirements: 90 of 102 (88.2%). Excludes 34 VERIFIED rows coded Unknown and 64 non-VERIFIED rows.
3. Payment requirements: 47 of 113 (41.6%), combining Yes and Partially coding. Excludes 23 VERIFIED rows coded Unknown and 64 non-VERIFIED rows.
4. Court access: 3 of 35 Yes (8.6%) and 32 of 35 Partially, among VERIFIED court rows with known access coding.
5. Property access: 24 of 29 Yes (82.8%), among VERIFIED property rows with known access coding; 2 VERIFIED property rows coded Unknown are excluded.
6. Harris County is used solely as an illustrative boundary case between an online database and an official repository, based on the county clerk's published explanation of its real-property portal.
7. Utah Courts documented 8 of 8 Privacy Guidance Clarity Score components in statewide judicial guidance. Documentation clarity only.
8. Dataset B was not hash-selected. Only Dataset C used outcome-blind SHA-256 ordering within frozen quotas, drawn from the frozen Dataset B pool.
9. All Dataset A state rows remained NEEDS REVIEW for publication clearance; no state prevalence figure is published.
10. Coding reflects official documentation available at the time of research. Government systems and guidance can change after review.