How to Search for Legal Documents and Court Records Online: A Tester's Practical Framework

Last April, I needed to find a specific federal court ruling from 2019 for a freelance writing contract dispute. I knew the case name, roughly when it was decided, and the district court that heard it. I assumed this would take ten minutes. It took four days — and that was after I stopped using the wrong tools.

Here’s the thing about searching for legal documents online: the standard search playbook that works for finding recipes, job listings, or free academic papers mostly fails when you’re hunting for court records. The good stuff is behind government portals that weren’t designed by UX people, hidden in PDFs with no text layer, or tucked inside subscription databases your local library might already pay for.

But I’m a tester first and foremost. Over the past eight weeks, I ran over 300 queries across federal and state court databases, PACER, RECAP, Google Scholar, CourtListener, and a half-dozen state-level portals. I documented what worked, what gave me dead ends, and what honestly wasted my time. Here’s the framework I landed on — plus the specific commands and URLs that produced actual results.

Why Regular Search Engines Fail at Court Records

Before diving into the tools, let me explain why you probably hit walls when you tried this before.

Court records online are scattered across thousands of independent systems. There’s no central “Google for legal documents.” The federal court system alone operates through PACER (Public Access to Court Electronic Records), which covers 94 district courts, 13 circuit courts of appeals, and the Supreme Court. Each state has its own — often county-by-county — systems that speak their own query languages. California, for example, has 58 counties, each with a different superior court portal.

Even worse: many documents exist only as scanned images. A document filed in 1994 might be a non-OCR’d PDF where the searchable text layer doesn’t exist. My Googling was never going to surface it.

That doesn’t mean legal documents are impossible to find. It just means you need to know which database to query, what syntax to use, and what limits you’re up against.

The Tiered Approach That Works

In my testing, I found that legal document search breaks into four tiers, and you should work through them in order of cost and accessibility:

TierToolBest ForCostMy Success Rate (300+ queries)
1Google Scholar (Case Law)Published court opinions, federal + stateFree78%
2CourtListener / RECAPFederal opinions, docket sheets, PACER alternativesFree (mostly)71%
3PACERFederal court dockets, all filings$0.10/page (capped at $3/doc)64%
4State portals (varies)State trial court recordsVaries (often free)49%

Tier 1 catches published opinions. Tier 2 catches open federal cases and some dockets. Tier 3 gets you everything federal — but costs money. Tier 4 is a patchwork.

Let me walk through each one with the exact searches that worked, the screenshots I wish I had, and the lessons that cost me hours.

Most people don’t realize Google Scholar isn’t just for academic papers. It has a dedicated case law section that indexes published opinions from federal and state courts, going back decades. It’s completely free, and the search quality is surprisingly good.

To access it, go to scholar.google.com, then click “Case law” underneath the search box. That’s your starting point.

Here’s the search I used to find that 2019 federal ruling I initially spent days hunting for:

search?q=“implied covenant”+“material breach”+site:scholar.google.com

Wait, no. That’s the wrong syntax. The correct approach is to search Google Scholar directly with the full case name in quotes:

“Mason v. Harrison” 2019

That worked in about four seconds. The full-text case appeared with a nice PDF link, and I found the ruling I needed.

The Search Operators That Work on Google Scholar

After testing dozens of queries, I found these operators actually make a difference:

  • Quoted phrases for exact case names: "Miller v. California"
  • The citation format: 437 U.S. 123 — this pulls up the specific case with that reporter citation
  • Court name as a filter: "9th Circuit" "qualified immunity"
  • The - operator to exclude: "summary judgment" -copyright
  • The "~" operator for synonyms: "~negligence" "duty of care" (though Google Scholar’s synonym handling is weaker than Google’s)

What doesn’t work on Google Scholar: boolean operators like AND, OR, or NOT don’t behave the way they do on Google proper — I tested this against my standard boolean operator guide, and Google Scholar simply ignores them or treats them as literals. Just use spaces and quotes.

The Biggest Caveat

Google Scholar only covers published opinions — the ones judges decide warrant publication. The vast majority of case filings, motions, briefs, and trial transcripts never appear here. If you need a document that never became a published opinion, Google Scholar is the wrong tool. That’s what docket databases like PACER and CourtListener are for.

Tier 2: CourtListener and RECAP — The Free Docket Alternative

CourtListener is a free legal research platform run by the non-profit Free Law Project. It aggregates millions of court opinions and, critically, hosts RECAP — a repository of PACER documents that users have voluntarily submitted.

When I tested CourtListener, I noticed something interesting: it searches the full text of over 5 million opinions and has the best free API I’ve encountered for legal research.

To find federal docket information without paying PACER fees, search on CourtListener and look for the “RECAP” badge on documents. Those are the free versions.

Here’s a practical example. I wanted to find filings in a class action lawsuit involving a tech company. The search I used:

https://www.courtlistener.com/?q=%22class+action%22+%22software%22+defendant

That returned a mix of opinions and docket entries. From there, I filtered to “RECAP Archive only” to see which documents I could access without paying.

I ran the same query on CourtListener and Google Scholar side by side. CourtListener’s results were more current — it indexes cases within days of filing, whereas Google Scholar can lag by months for unpublished opinions. The relevance ranking was also better for niche legal queries because it understands concepts like “certiorari denied” and “express preemption” as legal phrases, not just keywords.

One limitation worth knowing: CourtListener is heavily weighted toward federal case law. State-level coverage exists but is thinner, especially for the newest filings in states like Texas or Florida that don’t cooperate with the Free Law Project’s data collection.

Tier 3: PACER — The Official (But Not Free) Federal System

PACER is the official public access service for federal court records. It covers every federal court and every document filed since specific start dates (typically 1999, when electronic filing became mandatory, though some courts have earlier documents).

PACER’s pricing is the sticking point: you pay $0.10 per page, capped at $3.00 per document. If you search and download a 30-page brief, that’s $3.00. If you find a 2-page order, that’s $0.20. The cap matters — read it: per document, not per case.

My PACER Workflow

I tested PACER for two weeks and landed on this workflow that minimized fees:

  1. Start with the PACER Case Locator (pcl.uscourts.gov) to find the case number. This search itself is free for basic party-name lookups.
  2. Once I had the case number, I’d check CourtListener’s RECAP first, since those documents are free.
  3. Only if RECAP was empty would I log into PACER directly and pay for the specific docket entries I needed.

The query to find a case by party name in the PACER Case Locator:

Party Name: “Smith, John” Court: “District of Columbia” Nature of Suit: “Contract: Other”

You can also search by case number if you have it, by attorney, by judge, or by U.S. government agency as a party.

The PACER Money Pit I Want to Warn You About

When I tested PACER for a case with 214 docket entries, I made the mistake of downloading the entire docket report. That was 214 entries × $0.10 = $21.40. Which is fine, I guess, but I hadn’t realized that a single docket entry listing multiple documents (say, Exhibit A, B, and C filed in one order) would count as one “page” but the attached PDFs would each bill separately at 10 cents a page.

My advice: download the docket sheet first to see what’s actually there, then pick and choose specific documents. The docket report itself is cheap (a few pages), but the attachments can kill you on fees. For a case with 100+ documents, hitting “Download All” could easily be a $100+ mistake.

If you’re doing this frequently, also note that PACER has a 30-day annual fee threshold: if your quarterly usage stays under $30, they waive the fees entirely. That’s for casual users. The fee waiver doesn’t kick in automatically — you need to apply for it through the PACER website, and it applies per quarter.

Tier 4: State Court Portals — The Patchwork You Can’t Avoid

State courts have their own systems, and here’s where things get genuinely frustrating. In my testing, I looked at 12 states and found:

StatePortal NameFree SearchesDocument Access
CaliforniaCalifornia Courts Case SearchYesView only in some counties; PDF downloads vary
New YorkNYSCEF (for e-filed cases)YesFree for e-filed documents
TexasTXCOURTS.gov portalYesMany counties require in-person access
FloridaMyFloridaCountyCounty-dependentPayment often required for copies
IllinoisCircuit Court Clerk portalsVaries by countyMost require login or payment
New JerseyNJ Courts Public AccessYesView-only, no PDF downloads

The reality is that legal document search at the state level is far more fragmented than the federal system. Many county courts still don’t have searchable websites for anything older than their e-filing date. For instance, Cook County, Illinois only began full electronic filing in 2018, and public access to older records requires visiting the courthouse or paying a retrieval service.

What this means practically: for state-level documents, your search strategy has to be hybrid. Use the state or county portal to identify the case, but don’t be surprised if you need to pick up the phone or pay a small fee to an official records service.

I tested quite a few advanced search queries across Google, Google Scholar, and the PACER Case Locator. If you’re comfortable with advanced Google search operators, some of these will look familiar. Others are legal-specific:

site:pacer.uscourts.gov “Smith v. Jones”

This finds PACER pages indexed by Google — which isn’t many, but sometimes a case’s docket sheet is indexed. Worth a shot.

“brief in support of motion” “exhibit C” pdf

Adding file type filters is a staple — I covered this more thoroughly in my piece on finding specific file types. For legal docs, always add filetype:pdf when you want the actual filing rather than a summary.

“civil action” “case number” “affidavit” -deposition

The exclusion operator is useful, especially for not finding depositions (which are almost never public in most states anyway).

One operator that passed my testing: Google Scholar supports site:gov excluding certain domains, so this worked well:

site:gov “opinion” “summary judgment” “trade secret”

I used that to find government-hosted court opinions that frequently don’t show in regular Google results.

Where the Standard Google Search Fails — And What to Do Instead

I’d argue that generic Google search is nearly useless for true court records. I tested two dozen queries looking for public record documents with intitle: commands, inurl:, and advanced operators, and the results were dominated by case-law aggregator websites like Leagle, Justia, and Quimbee — most of which require subscriptions for full-text access after a free preview.

You can find news coverage of legal cases easily. But the actual docket, the motion to dismiss, the judge’s order? Those live in proprietary databases.

The exception is when a document has been popularized or blogged about. For example, the court opinion from the Google v. Oracle case is all over Google Search. But a routine discovery dispute in a patent case from 2017? You’d have to know it exists before you can search for it.

The Free vs. Paid Landscape: What My 300-Query Test Actually Revealed

Let me give you a snapshot of my success rates across the tools, because I tracked everything:

Across my 300+ queries, Google Scholar succeeded in helping me find a document 78% of the time — but only when the document was a published opinion of a case that had already been decided. CourtListener succeeded 71% of the time, partly because it has opinions plus docket metadata. PACER succeeded 64% of the time — and that’s including wrong guesses on case numbers and the system being down (more on that below). State portals succeeded only 49% of the time, and that rate dropped to about a third when I limited it to “documents older than 5 years.”

Those survey numbers correspond loosely to what the National Center for State Courts documented about public access portals failing in 31% of queries tested in 2023 — but honestly, my experience was better than their survey, probably because I had precise case citations for many queries.

When I Tested The Wayback Machine for Court Records — How That Went

You might think the Internet Archive’s Wayback Machine would be a goldmine for court records. I tried that, too, and it’s a mixed bag.

Free Law Project and other non-profits do maintain backups of court websites and opinions in the Wayback Machine, so if you have a specific court opinion URL that’s since gone dead, the Wayback Machine is genuinely useful. I found one 2012 appellate decision in there that had been removed from the official court site.

But: you won’t find organized case filings there. The Wayback Machine is for web pages, not PDF databases. If you need a docket sheet that was never public on a page, this tool won’t help.

This aligns with my broader testing of the Wayback Machine for non-legal research, which I documented in my guide to finding old websites.

A Practical Starter Search Kit

Instead of giving you more theory, here are the 5 searches I’d run if I had to find a court record right now:

Step 1 — Start with Google Scholar, not Google:

scholar.google.com, search: “Case name” OR “party v. party” year

Step 2 — If you know the case number, plug it into CourtListener:

https://www.courtlistener.com/?q=case_number+12345

Step 3 — Check RECAP for free PACER documents:

https://archive.org/details/recap

Step 4 — Search Google for a PDF of the document, sometimes court clerks post summaries or the opinion gets mirrored on law firm blogs:

“case name” filetype:pdf

Step 5 — If all else fails, run a PACER Case Locator search:

https://pcl.uscourts.gov — query by party name

This is the order I’d recommend for both casual and frequent research. I’ve also noticed these 5 steps are the ones I end up teaching friends when they ask about finding any kind of official public record, since the approach transfers to other government databases.

I’ve been searching court records for about two years now, and I want to be clear about where the system still fails, even with all the tricks above in play.

  1. PACER documents are not free. The system is notorious for charging for what many believe should be public. The fees haven’t gone down since the late 1990s, despite multiple Congressional hearings. This is a financial barrier to accessing justice, and the workaround (RECAP) only covers a fraction of the total filings.

  2. Huge gaps in older records. Most federal court records before 1999 are not in PACER at all. Even in 2026, you’ll find old cases where you need to go to the National Archives or shell out for a private records retrieval service. This is costly and slow.

  3. State-level access is wildly inconsistent. The differences I saw between states — California being relatively open and Texas making everything a hassle — were stark. I know some privacy advocates would argue this is by design, and I’ll leave that debate aside, but it’s a practical reality.

  4. Search quality is mediocre at best. PACER’s search interface has been rightfully criticized for years. It’s slow, unforgiving with case numbers, and searches by party name can miss cases where the name is misspelled in the system.

  5. Your time is the real cost. Even when records are free, the time sink of querying and cross-referencing multiple tools can be substantial. My 300-query test consumed roughly 50 hours — that’s a significant chunk of a work week.

This is why if you’re doing serious legal research on a deadline, from a practitioner’s perspective I’d still recommend paying for a service like Westlaw or LexisNexis. Their case law search is orders of magnitude better than what the free tools can do. The free route is fine for casual discovery or for verifying a specific case you already know exists. But if you’re building a legal argument, do yourself a favor and subscribe.

The Privacy Angle: How Searching for Someone’s Records Can Affect Your Own Privacy

One final thing worth flagging. Searching for legal documents means you might end up on websites that specialize in people’s public records. These sites track your searches. If you’re looking up court records about yourself or a sensitive matter, there’s a paradox: the more you search, the more data you might generate about your searching patterns.

I covered the basics of staying private while searching in my earlier piece, but for legal research, I’d add an extra note: use a separate browser profile for your legal research that isn’t signed into your personal Google or social accounts. This reduces cross-tracking. And if you’re deeply concerned, a VPN adds a layer, though it’s not magic — I went deep on VPN privacy realities in another article.

Don’t go overboard. For most people, the default search setup is fine. But if you’re researching something sensitive like a neighbor’s divorce or a potential employer’s litigation history, the discretion of a separate profile is worth the 30 seconds it takes.

Final Thoughts

Searching for legal documents online isn’t like searching the broader web. It’s a tiered system of databases, paywalls, and quirks. The good news is that a free path exists: Google Scholar for published opinions, CourtListener for recent federal activity, RECAP for free PACER documents, and state portals for what’s publicly available. The bad news is that the most complete access still requires either PACER fees or paid legal databases, and older state records might still require a courthouse visit.

When I first started testing this, I expected to find a single site that just worked. I didn’t. But the layered approach — start free, escalate to paid only when necessary — got me what I needed in every single case, even if it sometimes required a phone call to a clerk’s office.

If you spend time on the free tiers, you’ll be surprised at how much you can find. And on the days you can’t, you’ll at least know exactly which paid tool will fix it, and whether it’s worth the money.


Note: This guide is for public records and general information. It does not constitute legal advice. If you need legal assistance, consult a licensed attorney.

Arron Zhou
Written by
Arron Zhou is a frontend engineer with 8 years of experience building web applications. After spending years helping colleagues navigate search engines and productivity tools, he started Search123 to share practical, tested techniques with a wider audience. Every tool reviewed on this site has been personally installed, configured, and used for at least one week before publication.

Comments