Expungement orders the government to seal or destroy its copy of your record. It does not order UniCourt, Justia, CourtListener or a news site to delete theirs, because those companies were never parties to your case. That is the gap. Closing it takes a second process, aimed at each republisher and then at Google, and it is separate from anything the judge signed.
Why is my expunged record still showing on Google?
Because the order reached the government, and the copy on Google belongs to a private company. When you were arrested or charged, the docket was a public record. Commercial aggregators crawl those dockets continuously and keep their own copy on their own servers. Your expungement told the court, the prosecutor and the state repository to seal their file. It said nothing to the company that copied that file eighteen months earlier.
Google sits one step further out. It is not indexing the court’s record. It is indexing the aggregator’s page about the court’s record. Two layers stand between your order and the search result, and each has to be addressed on its own terms. That is the same structure behind removing court records from Google generally, and it is why one case usually produces six to twelve URLs rather than one.
Who does an expungement order actually bind?
The parties and the agencies it names: normally the court clerk, the prosecuting office, the arresting agency and the state criminal history repository. A private publisher that was never served and never appeared is a non-party, and a court generally cannot direct a non-party through an order it had no chance to contest.
There is a constitutional layer under that. In Cox Broadcasting Corp. v. Cohn, 420 U.S. 469 (1975), the Supreme Court held that a state may not impose sanctions on the accurate publication of information obtained from judicial records open to public inspection. The Second Circuit applied the same logic to an erasure statute in Martin v. Hearst Corp., 777 F.3d 546 (2d Cir. 2015), holding that Connecticut’s erasure statute did not require a newspaper to delete a truthful report of an arrest. The statute, the court wrote, “creates legal fictions, but it does not and cannot undo historical facts or convert once-true facts into falsehoods.”
None of that makes the order useless. Its power is persuasive rather than compulsory: most republishers act voluntarily when shown a certified order, and almost none act without one.
Do any state laws force a private site to take it down?
Several states penalize disclosure of a sealed or expunged record, but the statutes are narrower than people expect and none is a self-executing takedown. Two that are frequently cited:
- California. Penal Code section 851.92(c) attaches a civil penalty of $500 to $2,500 per violation for disseminating information about a sealed arrest. Enforcement runs through a city attorney, a district attorney or the Attorney General, not through you.
- New Jersey. N.J.S.A. 2C:52-30 makes it a disorderly persons offense to reveal the existence of an expunged arrest or conviction while knowing it was expunged. The maximum fine is $200.
These vary enormously by state, they interact with the constitutional problem above, and whether one reaches your facts is a question for your own lawyer.
Where is my case actually sitting, and what does each operator require?
Each republisher has its own policy, documentation standard, and definition of what “removed” means. Those differences decide the outcome. Every entry below comes from the operator’s own published policy.
| Operator | What it holds | Its stated route | What it requires | What “removed” means there |
|---|---|---|---|---|
| CourtListener (Free Law Project) | Federal dockets and opinions, including RECAP filings | Blocks search engines with a robots meta tag or X-Robots-Tag header on request. Will not remove a public document without a court order. | A request through its contact page listing exact URLs. A court order for deletion, redaction or anonymization. | Page stays live but carries a noindex instruction and falls out of Google. |
| Justia Dockets | Federal dockets, opinions and case filings | Will “consider written requests to disallow search engines from processing or scanning certain case filings and opinions using the robots.txt protocol.” | A written request through support.justia.com. | A robots.txt disallow, which is not noindex. See below. |
| UniCourt | State and federal dockets, party and attorney records | A redaction request route for records appearing in search engines. States that public court records are not “personal information” for CCPA deletion. | Its redaction request form. Responses within 45 days, extendable to 90. | Redaction of the published record, not deletion of the page. |
| PacerMonitor | Federal dockets and filings | No removal or de-indexing process published. | Undocumented. Only stated contact is support@pacermonitor.com. | Case by case, no published standard. |
| Trellis | State trial court dockets and rulings | No removal, redaction or opt-out process stated. Its privacy policy addresses account preferences only. | Undocumented. | Undocumented. |
| Google Search | The index pointing at all of the above | Legal removals, personal information removals in listed categories, refresh of outdated content. | Varies by route. Legal notices are strongest and most exposed. | The URL leaves Google. The source page is untouched. |
Why does a robots.txt block not get you off Google?
This is the trap that costs people months. A robots.txt disallow tells Google not to crawl a URL. It does not tell Google to drop a URL it already has. Google documents the consequence: “If the page is blocked by a robots.txt file or the crawler can’t access the page, the crawler will never see the noindex rule, and the page can still appear in search results.”
So blocking an already-indexed docket page can leave a stripped listing in place for a long time, because Google can no longer crawl it to learn it should go. A noindex rule or X-Robots-Tag header works precisely because the page stays crawlable and Google reads the instruction. So when an operator says it has “blocked the page from search engines,” ask which method. If the answer is robots.txt, expect the listing to linger and plan a second step at Google.
What is the fix sequence, step by step?
Order matters more than effort. Doing step six before step three is the most common reason a request gets denied and has to be started over.
- Get the certified order and read what it says. Sealed, expunged, dismissed and vacated prove different things to an operator. Send the certified copy, not a docket printout.
- Inventory every URL. Search your name in quotes, your name plus the case number, and your name plus the county. Build the full list before contacting anyone.
- Go back to the court first. If the clerk has not pulled the docket from the public portal, aggregators will re-crawl it and the case returns after you have cleaned it up. Close the source first.
- Write to each operator separately, with exact URLs and the certified order attached. Use the route the operator publishes, not a general contact form.
- Ask for noindex, not robots.txt. Say it in the request. Operators that offer both often use whichever you name.
- Only then use Google’s tools. Its published denial reasons for outdated content requests include “The text or image that you described in your refresh outdated content request is still present on the live page.” That is Google telling you the source was never fixed.
- Escalate to a legal removal request where an order supports it. Know the trade-off first: Google states it “may send a copy of each legal notice we receive to the Lumen project for publication,” without the submitter’s contact information.
- Re-check at 30, 60 and 90 days. Scrapers and syndication partners surface later than the originals. A case that looks clean in week two often is not in month three.
What does this cost?
Link removal is priced per URL, typically $1,000 to $2,000 per link, quoted after an eligibility check. No retainer, no monthly minimum.
What surprises people is the link count, not the per-link price. One expungement can mean six to twelve removals, because CourtListener, Justia, UniCourt, PacerMonitor, Trellis and DocketBird each hold their own copy of the same docket, with news coverage on top. That arithmetic is what a single flat price for “an expungement cleanup” leaves out. Criminal record removal and court record removal are scoped the same way.
Suppression is a different product. It is quoted per engagement, costs more than a single removal, and exists only as the fallback when removal is genuinely unavailable. Try removal first.
What this does not do
- It does not delete the court record or the source page. Removal takes a URL out of search results. The docket and the publisher’s page continue to exist.
- It does not defeat a background check. Screeners that pull from courts, state repositories and county clerks never touch Google. Under 15 U.S.C. section 1681c(a)(2), consumer reporting agencies may not report arrest records predating the report by more than seven years, but subsection (b) lifts that limit for credit transactions of $150,000 or more and jobs paying $75,000 or more a year. Where a report is inaccurate, the FCRA dispute process is a separate route with its own rights.
- Any disclosure obligation you have survives a removal entirely. If an employer, licensing board, regulator, lender, counterparty or court is entitled to ask, your state’s expungement law and your own counsel govern what you must say. Nothing here is a way around that, and we do not take work framed that way.
- Removed content can come back. Scrapers, mirrors and syndication feeds resurface material months after the original is gone.
- Some things are not removable at all. Accurate news reporting, valid unexpunged convictions and sex offender registry listings stay. Anyone who tells you otherwise is selling something they cannot deliver.
Frequently asked questions
Does Google honor an expungement order on its own?
Not automatically. Google’s personal information removal policy covers categories such as government ID numbers, financial account numbers, medical records and contact details. Court records are not among them. An expungement order supports a legal removal request and a request to the republisher, but it is not something Google acts on by itself.
How long does this take?
It depends on the operator, and they differ by an order of magnitude. UniCourt commits to answering a deletion request within 45 days, extendable to 90. CourtListener warns that even after it de-indexes, search engines “may not remove your case from their results for many months, if at all.” Plan in months.
Should I pay a site that offers to remove my record for a fee?
No. Paying a site to take down its own page rewards the business model that put it there, marks you as someone who pays, and does nothing about the other copies. Several states have moved against paid-removal arrangements in the mugshot context specifically.
Will my removal request itself become public?
Possibly, on the legal route. Google states it may send a copy of each legal notice it receives to the Lumen project for publication, minus the submitter’s contact information. That is a real consideration when the case is sensitive, and a reason to exhaust the operator-level routes first.
Start with the inventory
Before deciding what to spend, find out how many links the case produced and which have a real removal basis. That is what the Search Risk Audit does: an inventory of what is indexed under your name, the route available for each URL, and a per-link quote for the eligible ones. If a link has no honest basis for removal, we say so.
Researched and reviewed by the Search Zapper editorial team. Published September 9, 2026. We describe mechanics, not legal advice.