Delete Yourself From the Internet: A Global Guide

She found out the way most people do: someone told her. Her neighbor’s husband — someone she’d never given her phone number to — had texted her out of nowhere. He knew her address too. She hadn’t posted either online. She thought she was careful. She wasn’t.

What he had used was a people-search site. These services collect your name, address, phone, and relatives’ names from public records, voter rolls, and property databases. Then they sell access for $4.99 a month to anyone who wants it. No background check. No stated purpose. Just a search box.

This is what “deleting yourself from the internet” is actually about. It’s not mainly about scrubbing embarrassing tweets, though that matters too. It’s about removing the infrastructure that lets strangers locate you, contact you, and build a picture of your life without your knowledge.

The honest version of this guide is not “here’s how to disappear completely.” You can’t. But you can dramatically reduce what’s findable about you, by whom, and how quickly. This guide covers every major layer of your digital footprint, in order of impact. It’s direct about which layers you can fix and which you can only limit.

The short answer. You can’t fully delete yourself, but you can shrink your footprint layer by layer. Start with the highest-impact one: data brokers and people-search sites like Spokeo, Whitepages, and BeenVerified, plus upstream aggregators like Acxiom. Then handle Google search results, old social accounts, and marketing databases. Public records, breach data, and AI-training datasets can only be limited, not erased. Removed listings reappear as source records refresh, so treat this as recurring maintenance.

Jump to a section:

Why people do this

The reasons vary, and the priority order matters.

Safety-critical cases — domestic abuse survivors, stalking victims, people who’ve received threats — need to move fast on the highest-exposure items first. People-search sites and data brokers are the primary threat surface here, not Google’s index. A stalker doesn’t need your name to appear in a news article; they need your current address.

Reputation management often follows something specific: a difficult breakup, an embarrassing incident from years ago, an old social media post surfacing at a bad time. These cases center on search engine visibility — what appears when an employer, date, or journalist searches your name.

Proactive privacy is the quieter but increasingly common reason. People who’ve watched one too many data breach notifications arrive in their inbox, or who simply don’t want their income estimates and family tree accessible to strangers.

Pre-employment sweeps are becoming standard among people who understand how hiring works. Knowing what an employer will find before they find it gives you time to do something about it.

Layer 1: What Google shows about you

Start with Google search. Open an incognito window, search your full name, and add the city you live in. What appears on the first page is what most people see when they look you up. That’s your baseline.

A few common things you’ll find:

Social media profiles — LinkedIn, Facebook, Instagram, Twitter/X. These appear because the platforms are public and indexed by design.

People-search site listings — Spokeo, Whitepages, BeenVerified showing your address, phone, and relatives. These are served up because the sites want traffic, and Google happily indexes them.

Old news articles, forum posts, or comment threads with your name. These often turn up from accounts-on-websites you’ve long forgotten.

Professional profiles — your company’s team page, a conference speaker bio, a directory listing.

Removing content from Google search

Google will de-index — meaning remove from search results — certain categories of content upon request:

Your home address, phone number, email address when appearing on non-authoritative sites
Intimate images published without your consent
Doxing content (personal information posted to facilitate harassment)
Minors’ data in some jurisdictions

Go to support.google.com/websearch/troubleshooter/9685456 for the removal request form. Be specific about the URLs. Google typically processes these within a few days to a few weeks.

What this does not do: it removes the result from Google’s index, not the source page. The original site still has the content. Anyone using Bing, DuckDuckGo, or a direct URL still finds it. Google de-indexing is a visibility reduction, not a deletion.

Bing has a similar removal request process at bing.com/webmaster/tools/contentremoval. DuckDuckGo relies on Bing’s index for most results, so removing from Bing typically removes from DuckDuckGo as well.

This layer is straightforward in principle and tedious in practice.

For active platforms you use, review your privacy settings. Most major platforms default to more public than they need to be. On Facebook, your friends list, past posts, and tagged photos can each be restricted separately. On Instagram, switching to a private account hides your posts from non-followers but keeps your profile name searchable. LinkedIn lets you control whether your profile appears in public search results — it’s under Settings > Visibility > Profile visibility off LinkedIn.

The harder problem is old accounts. Most people have profiles on five to fifteen platforms they no longer check: Tumblr, MySpace archives, old Reddit accounts, Flickr, forums from a decade ago. These persist and rank in search results precisely because the accounts are inactive and no one is managing them.

Go through them methodically:

Log in and delete the account entirely if the platform allows it (most do, buried in Settings > Account > Delete)
Where deletion isn’t offered, delete all content, change the name to something generic, remove the profile photo, and make the account as private as possible
For platforms where you’ve lost the login, use the password recovery flow first; if the email no longer exists, contact support with proof of identity

Account deletion on most major platforms takes 30 to 90 days to fully propagate through the platform’s systems and disappear from search engines. Don’t expect it to be instant.

Tagged photos are a separate problem. Photos others have posted of you, at events or in group shots, persist even if you delete your own account. Your options here are narrow. Request deletion from the person who posted it. Or ask the platform to remove it under its policies. In GDPR jurisdictions, you can also file a deletion request with the platform directly.

Layer 3: Data brokers

This is the layer that matters most for practical privacy, and the one most people underestimate.

There are roughly 4,000 data broker companies in the US alone. They collect your personal data from public records, retail purchase histories, loyalty programs, warranty cards, motor vehicle records, and dozens of other sources. Then they sell access to that profile to insurers, employers, advertisers, landlords, and anyone else willing to pay.

The information in these profiles is detailed. It’s not just your name and address. It’s your estimated income, household composition, health-condition categories, political affiliation, buying habits, vehicle, and sometimes your daily movement patterns inferred from your phone.

US data brokers

The major people-facing sites each have opt-out forms: Spokeo, BeenVerified, Whitepages, MyLife, Intelius, TruthFinder, Radaris, and FastPeopleSearch. The opt-out process is voluntary under federal law for most of them. California’s CCPA makes it somewhat enforceable if you’re a California resident. Each site has a slightly different process: submit your name and location, confirm via email, then wait 5 to 30 days.

For the full walkthrough of each major people-search site’s opt-out URL and process, see our people-search sites removal guide.

For the upstream wholesale data aggregators — Acxiom, Oracle Data Cloud, Epsilon, LexisNexis — the opt-out process is different and less user-friendly. Our complete US data broker removal guide covers both tiers.

California’s DELETE Act (SB 362) covers all registered data brokers, roughly 500 in California’s registry. It requires them to honor deletion requests submitted through a single state portal called DROP. The portal opened to consumers on January 1, 2026, and registered brokers must start processing requests from August 1, 2026. If you’re a California resident, one filing reaches all those registered brokers at once. It doesn’t cover unregistered brokers or companies outside the registry. For the full mechanics, see how to file a DELETE Act request.

Timeline: Each broker typically processes opt-out requests within 30 to 45 days. Your information will reappear. Data brokers re-acquire records from new source data: updated voter rolls, new property transactions, fresh social media scrapes. Most privacy researchers recommend reviewing and re-submitting every 6 to 12 months.

EU data brokers

In the EU, you have a legal right — not a voluntary courtesy — to demand deletion under GDPR Article 17. Data brokers must comply within 30 days or provide a written legal basis for refusing. The right applies regardless of where the broker is headquartered, if they’re processing data about EU residents.

The process is direct. Find the company’s Data Protection Officer contact, required by GDPR Article 37 for large-scale processors and published in their privacy policy. Send a formal Article 17 erasure request citing your name and the data you want deleted, and note the date. If they don’t comply within 30 days, that’s a reportable violation.

The major EU brokers include Acxiom EU, Experian, Epsilon France, Bisnode, Creditreform, and country-specific directory sites like 118.nl and Das Telefonbuch. See our EU data broker removal guide for specific DPO contact addresses and templates.

India data brokers

India’s Digital Personal Data Protection Act (DPDPA), passed in 2023, establishes a right to erasure. India notified the DPDP Rules and stood up the Data Protection Board in late 2025. But 2026 is a phased “soft enforcement” period, with the heaviest penalties for core obligations arriving by mid-2027. In practical terms, your erasure rights exist and the Board is live, but the formal mechanism to compel a non-compliant broker is still ramping up.

What you can do now: send erasure requests to Indian data broker companies directly using the DPDPA’s erasure right as your basis. Many companies have designated grievance officers — distinct from the EU’s DPO model — whose contact details they’re required to publish. The practical process for Indian data brokers is covered in our DPDPA rights guide.

Removing yourself from Indian internet directories and aggregation services is its own challenge. The removing yourself from the Indian internet guide goes deeper on country-specific sites.

Layer 4: Public records

This is where the honest part of this guide matters most. A large category of information about you is held in government databases — and government records are, by design, public. Your ability to remove this data is limited.

Court records — civil suits, criminal cases, traffic violations, evictions — are public records in virtually every jurisdiction. They’re indexed. They appear in people-search profiles. They feed background check services. In the US, expungement seals or destroys the court’s record of certain criminal cases, but the laws vary enormously by state. Expungement also doesn’t automatically remove data that a private data broker collected beforehand. You’d need to submit a separate request to each broker, citing the expungement.

Voter registration records are public in most US states. Some states (California, New York, Massachusetts) restrict access to party-affiliated groups or researchers. Others (Florida, Ohio, Michigan) make the full database available online. You cannot typically opt out of voter registration data while remaining a registered voter — the public nature of the record is a feature of election transparency. What you can sometimes do is apply for a confidential voter registration under your state’s address confidentiality program (ACP), if you qualify. Eligibility is typically limited to domestic abuse survivors, law enforcement, or public officials.

Property records are public by design. Property ownership and sale prices are recorded in county assessor databases that are freely accessible. You cannot remove them. What you can do is reduce their utility. Many people-search sites will eventually remove or suppress your listing if you submit opt-out requests, even though they pull from public data.

Business filings and licenses — if you’ve registered a business, obtained a professional license, or sat on a board, that’s in a state or federal database and is public.

Layer 5: AI training datasets

This is a newer layer, and the options are more limited than most coverage suggests.

Say your photo, writing, or personal data appeared on a website that was included in an AI training dataset. That data may have been incorporated into a model’s weights. The model doesn’t store your data like a database — it encodes patterns from it. There’s no straightforward way to “delete” information from a trained model. What you can sometimes do is request removal from the dataset itself before training, or request that a company not use your data in future training.

Stable Diffusion / Stability AI operates a removal request tool at haveibeentrained.com, built on the LAION dataset. There you can search for your images and flag them for exclusion from future training runs. This removes your data from the next version of the dataset, not from models already trained on it.

LAION (Large-scale Artificial Intelligence Open Network) also accepts removal requests through their contact form. LAION-5B was the primary dataset behind many open-source image models; if your images were scraped from the public web before 2022, they’re likely in it.

For text-based AI models from OpenAI, Google, Meta, and Anthropic, your options are limited. You can opt out of your future interactions being used for training, covered in our AI training opt-out guide. There’s no mechanism to retroactively remove your public writing from what these models were trained on.

Layer 6: Marketing and email databases

Marketing data is the least dangerous layer but the most pervasive nuisance. Your email and phone are almost certainly in hundreds of commercial databases used for direct marketing.

CAN-SPAM (US): Any commercial email must include an unsubscribe link. Using it obligates the sender to remove you within 10 business days. The catch is that each sender is separate — unsubscribing from one doesn’t affect others.

GDPR (EU) Article 21: You have the right to object to processing for direct marketing purposes. Unlike Article 17 erasure, the company can’t override an Article 21 objection with “legitimate interests” — once you object to direct marketing, they must stop. This includes profiling for marketing. Send a written objection, keep a record.

The Wayback Machine (archive.org) archives web pages, including pages that may have contained your personal information years ago. The Wayback Machine honors robots.txt exclusions — if a site’s current robots.txt blocks archival, the Wayback Machine will typically remove archived versions of that site on request. For personal pages, you can submit a removal request at help.archive.org/help/change-my-privacy-settings-in-the-wayback-machine/. This takes time and is not guaranteed for all archived content.

What cannot be deleted

Be clear on these. A guide that glosses over the limits wastes your time and creates false confidence.

Published news articles about you will not be removed because you don’t like them. Google can de-index them, making them harder to find. But the article stays live on the publisher’s site, and any other search engine can still surface it. Journalists and publishers have their own First Amendment protections. Your options are narrow. You can request corrections if the article contains factual errors. In extreme cases involving old content, you can file a “right to be forgotten” de-listing request with Google under EU law. That right applies only in EU searches and is decided case by case.

Breach databases once they’re in the wild are not retrievable. Say your email and password appeared in the Dropbox breach of 2012 or the LinkedIn breach of 2021. That data is now in circulation on dark web forums and in credential-stuffing lists. It cannot be recalled. HaveIBeenPwned deliberately retains breach notification data. It keeps not your password, but the fact that your account was in a breach. The service does this because it functions as a security tool for defenders. Your email cannot be removed from breach notifications, and this is a policy choice HIBP maintains intentionally.

Data already sold to third parties creates a downstream problem GDPR only partially solves. When you send an Article 17 erasure request to a data controller, they’re legally required to inform downstream processors who’ve received your data. In practice, verification is nearly impossible. The controller deletes their copy; whether the buyer of their list also deletes it is difficult to audit. Your data may persist in derivative datasets for years.

Inferred data is in a category of its own. Your actual income figure — if you disclosed it somewhere — is a record that can be deleted. Your estimated income, inferred from your zip code, vehicle registration, and purchasing patterns, is a derived computation. It’s not “your data” in the same way. You didn’t disclose it; it was calculated. Whether erasure requests cover inferred data is currently contested in several jurisdictions.

Professional licensing and credentialing databases are run by state agencies and professional bodies. Your medical license, law license, securities registration, or contractor license is public record by design. It cannot be removed.

The practical roadmap

Given all of the above, here’s how to prioritize.

The most common mistake is spending an hour getting your Facebook privacy settings right while doing nothing about Spokeo. People-search sites are the highest-exposure, highest-impact target — they’re what scammers, stalkers, and low-effort investigators actually use. Start there.

Deletion services vs. manual removal is a real tradeoff. Services like DeleteMe, Privacy Bee, and Optery automate much of the broker opt-out process — submitting requests, monitoring for re-listing, and running periodic sweeps. They typically cost $100 to $200 a year. The value proposition is time, not completeness: no service covers all 4,000+ brokers. If you’re dealing with a safety-critical situation, get started on the highest-exposure sites yourself today while setting up a service for ongoing maintenance.

Priority checklist

Start here and work down. Each item is worth doing even if you stop before finishing the list.

Highest priority — do these first

Search your name in incognito mode to establish a baseline
Submit opt-outs to the top 10 people-search sites (Spokeo, BeenVerified, Whitepages, MyLife, Intelius, TruthFinder, Radaris, FastPeopleSearch, Peekyou, Instant Checkmate)
Submit opt-outs to upstream data aggregators (Acxiom, Oracle Data Cloud, Epsilon, LexisNexis)
If you’re a California resident, file through the DELETE Act (DROP) portal on the CPPA’s official site
If you’re an EU resident, send GDPR Article 17 erasure requests to EU brokers

High priority — do these within a week

Audit your active social media privacy settings (Facebook, Instagram, LinkedIn, Twitter/X)
Find and delete or deactivate old social media accounts you no longer use
Request removal of your home address and phone from Google Search using their removal tool
Opt out of AI training data collection at OpenAI, Google, Meta (see our AI training opt-out guide)

Medium priority — complete within a month

Check Wayback Machine for archived personal pages; submit removal request if needed
Review employer and professional directory listings; contact HR or the listing owner for removal
Unsubscribe from marketing emails using CAN-SPAM process; for EU, send Article 21 objection to marketing lists
Check HaveIBeenPwned to understand your breach exposure; change passwords on affected accounts

Ongoing maintenance — set a reminder

Re-check people-search sites every 6 months; re-submit opt-outs for any reappearances
Google yourself quarterly to catch new listings
Check for new data breaches involving your email on HaveIBeenPwned

The process isn’t fast and it isn’t permanent. Data re-aggregates. Brokers buy new source lists. Old accounts get re-scraped. What you’re building is not a clean slate — it’s a practice. The goal is to be a harder target than you are today, consistently, over time.

That matters. Because most people never start.