What Is Legal Data? A Buyer's Guide to Types, Sources and Cost

Short answer: Legal data is the text of the law and the records around it. That means statutes (laws passed by a legislature), regulations (rules written by government agencies), court decisions, and the books and articles lawyers write about them. This guide uses that meaning. If you came looking for law firm billing data or e-discovery data (the documents gathered for a lawsuit), this is the wrong page. If you build or buy a product that has to be right about the law, keep reading.

TL;DR

• Legal data comes in four kinds: enacted law and agency rules, court decisions and dockets, commentary, and contracts and business records.
• Governments publish most primary law for free. Commentary and some court records are licensed or charged per use.
• A legal data provider is a company or nonprofit that collects this material, cleans it up and serves it to you. Price depends on coverage, freshness and what you are allowed to do with it.
• Bad data fails in four ways: it is out of date, it is the wrong version, it is unofficial, or its licence forbids your use.
• Test any source by checking twenty citations you already know against the official text.

The four kinds of legal data

Lawyers call both of the first two kinds primary law, meaning the law itself and not writing about the law. Court decisions are primary law too. This guide splits them because they come from different places and are bought differently. Commentary and contracts only describe or apply the law. Our guide to the types of primary law in the US goes deeper on the first row.

If your product tells people what a rule requires, you need the enacted text first. Commentary helps people understand it, but it cannot replace the source.

Where legal data comes from

There are two kinds of source.

Government publishers write or issue the law. A legislature publishes its statutes. An agency publishes its rules. A court publishes its decisions. On the federal side, govinfo is a service of the Government Publishing Office, and it describes itself as free public access to official publications from all three branches of the federal government.

Aggregators gather material from many publishers and put it in one place. Some are nonprofits. CourtListener, run by the nonprofit Free Law Project, is an open source of case law. Some are companies that sell research tools or data feeds. Vaquill AI is one example: it serves US statutes, regulations and other primary law through an API (a way for software to request data).

Why use an aggregator at all? Because fifty states publish in fifty different ways. One posts web pages, another posts PDFs, a third posts word-processor files. An aggregator turns that mess into one consistent shape, so your product does not need fifty separate collection jobs.

What drives the cost

Five things set the price of a legal data provider:

1. Coverage. Federal law only, or all fifty states plus DC and territories.
2. Cleaning. Turning web pages and PDFs into structured text with reliable citations.
3. Freshness. How quickly a change at the source shows up in the data.
4. Rights. Whether the provider owns or licenses what it sells, and what it lets you do with it.
5. Your volume. A few lookups a day or millions.

Pricing usually follows one of these patterns:

• Free. Government sites and nonprofit sources.
• Per record. The federal court system's PACER charges $0.10 a page, with a cap of $3 per document, and waives the fee if you spend $30 or less in a quarter. PACER publishes the details. Those numbers are about to move. The courts have announced a rise to 12 cents a page and a $40 quarterly waiver starting January 1, 2027, according to a court notice.
• Subscription per user. Common for research tools used by people.
• Usage-based. You pay per request or buy credits.
• Enterprise licence. A negotiated contract, often with limits on resale and display.

Why federal court records cost money

Most government data is free, so the PACER meter surprises people. PACER stands for Public Access to Court Electronic Records. It dates to September 1988, when the Judicial Conference of the United States (the body that sets policy for the federal courts) approved it as an experiment. Early users dialed in by modem, pressing a phone receiver into a rubber cup to read docket sheets on a screen.

The fee is just as old. The judiciary says on its own site that it asked Congress for money in 1988, Congress did not provide it, and the courts were told to pay for the system with user fees. A federal law, as amended in 2002, lets the courts charge "only to the extent necessary," according to a Congressional Research Service summary. What counts as necessary has been argued ever since. In 2016 three nonprofits filed a class action, National Veterans Legal Services Program v. United States, saying the courts had collected and spent more fee money than the law allowed. In August 2020 the Federal Circuit (the appeals court that heard it) ruled that the law "limits PACER fees to the amount needed to cover expenses incurred in services providing public access to federal court electronic docketing information."

In their words

Both sides of the argument show up on Hacker News, a forum popular with developers. In June 2026, after the Electronic Frontier Foundation published a piece backing the Open Courts Act of 2026, a bill that would replace PACER and end its fees, a commenter who goes by tptacek gave a mixed verdict. Their comment: "Speaking as a nonlawyer who uses PACER kind of weirdly a lot, it's also not good; it's pretty archaic. But it's also very cheap."

The director of Free Law Project, posting as mlissner in August 2020, recalled that the group's expert testimony in the class action put the cost of running PACER at "half a ten thousandth of a penny per page, IIRC." The "IIRC" means "if I recall correctly," so treat it as a recollection and not a ruling. The court filing the director linked is the place to check it.

For a buyer the useful part is simpler. A per-page meter is cheap for a lawyer who pulls five filings and costly for a product that pulls thousands. Price court records at your real volume before you promise anything to a customer.

Always get the rights in writing. Ask whether you may show the data to your customers, store a copy, resell it, or use it to train an AI system.

What goes wrong

It is out of date. A legislature repeals a section (ends it) and your data still shows it as law. Ask the provider how often it checks each source and whether every record shows the date it was last checked. Our post on legal data freshness explains what to look for.

It is the wrong version. A dispute about something that happened in 2023 is judged under the law as it read in 2023, which may differ from today's text. Ask whether the source can show a section as it stood on a past date.

It is unofficial. Some well-known free sites are convenient but say plainly that they are not the legal record. The federal rules site eCFR is a good example. (CFR stands for the Code of Federal Regulations, the official collection of federal agency rules.) It describes itself as "an editorial compilation of CFR material and amendments published in the daily Federal Register," and says it is maintained as an informational resource. Read its own explanation. For a product that must be right, know which version you are quoting. Our note on where a statute came from shows how to check.

The licence forbids your use. A source can be free to read and still bar you from copying it into a product. Ownership claims over legal material have also ended up in court. In 2020 the US Supreme Court decided that the notes Georgia published under its state code could not be copyrighted. The opinion is online. The lesson for a buyer is simple. Do not assume. Read the terms for your exact use.

Here is the story behind that case. Georgia's official code comes with annotations, short notes under each law that summarize court decisions and attorney general opinions. A state commission made up mostly of legislators supervises them, and a LexisNexis division drafted them under contract. At the time of the ruling, a printed set cost $412. A nonprofit called Public.Resource.Org posted the whole thing online for free, and Georgia sued for copyright. Chief Justice Roberts wrote for the Court that the idea behind the rule is that "no one can own the law." The Court added that the rule does not reach works by people with no power to make law, such as court reporters. A lower court in the same fight had noted the oddity of the situation: "most official codes are not annotated and most annotated codes are not official." When you hear "the official code," ask whether the notes under it come with it, and who owns them.

How to pick a source

Work through these steps in order.

1. Decide what you must be right about. Statute and regulation text points to enacted law. Court outcomes point to court decisions. A plain explanation for a reader points to commentary.
2. Decide who sees the data. If only your staff read it, many sources work. If your customers see it or your software repeats it, the licence matters most.
3. Decide how current it must be. Today only, or also as of a past date.
4. Decide how much work you want. A raw file you maintain is cheap to start and costly to run. A service costs more per month and saves the upkeep.
5. Test twenty citations. Pick twenty laws you already know. Check each against the official text. Count the misses, and look at how the source shows its sources.
6. Get the licence in writing. Confirm display, storage, resale and AI training.

This is general information, not legal advice.

If you need US primary law (statutes, regulations and similar sources) as data you can build on, see the legal API.

This guide is part of US Law Data: The Complete Guide, a map of where US law comes from and how to use it.

FAQ

What is legal data in simple terms?

It is the text of the law and the records around it. That includes statutes, regulations, court decisions, and the commentary written about them. It does not usually mean law firm business data, such as billing records.

Is legal data free?

Much of it is. Governments publish statutes and regulations for free, and nonprofits publish a lot of case law. Fees show up for federal docket records, for commentary, and for services that clean the data, keep it current and support it.

What is a legal data provider?

A legal data provider collects law from government publishers, standardizes it and delivers it to you, usually by website, file download or API. Providers range from free nonprofit projects to large commercial publishers.

Is case law legal data?

Yes. Case law is the written decisions of courts. It is primary law, like statutes and regulations, but it has its own sources and its own rules about cost and reuse.

Can I use legal data to train an AI model?

It depends on the source and its licence. Government text is often open, while a provider's compiled collection may carry limits. Read the terms for training use before you start, and ask the provider in writing.

How do I know if a legal data source is accurate?

Test it. Take a set of citations you can check yourself, compare the text to the official publisher, and note how current each record is. A good source shows where each record came from and when it was last checked.

Priyansh Khodiyar

Co-Founder & CTO

Priyansh leads engineering and AI at Vaquill AI: the pipelines that pull statutes, regulations and court rules from every US jurisdiction's official publisher, and the REST API, MCP server and open dataset that serve them.