Short answer: Firecrawl's Government & Legal Index is a search index of US law that an AI assistant such as Claude can query. Each answer comes back as a short quote, its citation and a link to the publisher. Vaquill AI is one of its design partners, and our part is advice: what to put in, what to leave out, where to take it from, and how statutes and regulations should be laid out so a citation lands on the exact provision.
Disclosure: we build Vaquill AI, a US primary law API, and Firecrawl's page lists us under "Tested by our design partners". Everything below about the index comes from that public page as it read on 22 October 2026. What we say about legal data is our own view. This is general information, not legal advice.
TL;DR
- What it is: a US law index from Firecrawl, which it describes as "30M+ data points of US law", built so an assistant answers with a quote, a citation and a source link.
- Our role: design-partner advice on the scope and shape of the data. We do not run the index, and this post is not a description of how it works inside.
- The rule we push hardest: the source decides everything. An official page beats a copy of it.
- Statutes and regulations are different shapes. A statute is a numbered section in a code. A regulation also needs its agency, its edition date and the statute it rests on.
- Leaving things out is half the job. Bills, drafts and unsourced summaries do not belong next to law.
What the index is, from their page
Firecrawl describes the index as "30M+ data points of US law". Its page works through five sample questions, such as California's 2027 minimum wage and how much notice a Texas landlord must give before filing an eviction suit.
Each answer is a short quote. It names the type of document (news release, statute, agency, legal FAQ), the publisher's domain, and a citation where there is one, such as Tex. Prop. Code § 24.005(a).
A second feature follows citations. Legal text keeps pointing to other legal text, so each result lists the statutes and cases it cites. The page shows Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), citing 194 authorities.
What a design partner does
A design partner does not build the product. We sit where the choices get made and give an opinion from the data side. For this index, that means what to add, what to leave out, which sources to take it from, which details matter and which do not, and how statutes and regulations should be structured.
We build a US primary law API, so we have made those same choices ourselves. That is the only reason our opinion is worth asking for.
Five things we push on
1. Take it from the publisher
Many sites repeat the law with no date and no author. A copy of a copy can be wrong in a way you cannot see, so the index should point at the page the legislature or agency itself publishes. We wrote about why in how to search the US Code and trust the result.
Firecrawl's examples do this. Each answer names a publisher domain, such as a state legislature or the FTC. Size matters less than that: 30 million points from weak pages are worth less than a million from the right ones.
2. Say what kind of document it is
A statute binds, and so does a regulation, through an agency. Agency guidance usually does not, and a press release only describes a rule. We cover the gap in do agency guidance documents have the force of law.
An assistant that flattens these types will quote a news release as if it were the law. The label has to travel with the text, which is why we like that the sample answers on their page carry one.
3. Currency beats coverage
The useful question is not whether a document exists but whether it is in force, and as of when. One sample on their page quotes the FTC's own page saying the Noncompete Rule is not in effect after an August 2024 court order. A model that learned the rule's text from an older source would get that wrong.
Keeping a date on every document is dull work and it is the work that matters. Our own corpus carries a currency date for the same reason.
4. Statutes and regulations are different shapes
Both look like numbered sections, but a good index stores them differently.
| Statute | Regulation | |
|---|---|---|
| Where it lives | Code title, chapter, section | Agency, title, part, section |
| Example citation | 42 U.S.C. § 1983 | 17 C.F.R. 240.10b-5 |
| What changes it | A new Act, with its own effective date | A rulemaking, then a new yearly edition |
| What to keep with the text | Subsections, history notes, amendments | Issuing agency, edition date, the statute it rests on |
A citation should land on the exact provision, not the page that contains it. We go through the US Code side in how to search the US Code and the regulation side in the eCFR guide.
5. Leave out what is not law
A bill is a proposal, not a law, as we explain in bills versus laws. Add to that drafts, law-firm summaries, scraped copies and old editions presented as current. Each one looks plausible, and each puts a wrong answer one step from a right one.
Most of a government web page is also not law: menus, banners and photos. Cutting that out is a feature.
What a legal answer should carry
Our view of the path from source to answer. It is not a diagram of Firecrawl's system.
Check any answer yourself
An index like this is only as good as your habit of checking it. Five steps take about two minutes.
- Open the source link. Read the page, not just the quote.
- Read the label. A statute, a regulation and a news release carry different weight.
- Find the date. Look for when the text took effect and whether a later law changed it.
- Resolve the citation. It should lead to the exact section.
- Follow one cited authority. If the answer rests on a case or a statute, open that too.
Our guide on checking what Claude tells you about statutes covers the same steps in more detail.
An assistant answers a question about a state wage rule by quoting an agency press release. How should you read that answer?
FAQ
What is Firecrawl's Government & Legal Index?
It is a search index of US law from Firecrawl, which it describes as "30M+ data points of US law", designed for AI assistants. Each answer carries a quote, a citation where there is one, the publisher and a source link. You can see it at firecrawl.dev/gov-index.
What is a design partner?
A design partner is an outside team that gives a company input while a product is being shaped, before it is final. Here, our input is on the legal data: what goes in, what stays out, where it comes from and how it is structured.
Is Vaquill AI's data inside Firecrawl's index?
This post does not say. We advise on the shape of the index, and we describe only what their page shows. For the list of sources, ask Firecrawl.
Can I rely on an AI answer built from the index?
Use it as a starting point. The quote and the source link let you check it quickly, and you should, because a rule may have changed since the page was captured. This is general information, not legal advice.
Why do statutes and regulations need different structure?
A statute sits in a code and changes by a new Act. A regulation sits under an agency, is republished in yearly editions, and depends on a statute that authorizes it. If the index stores both the same way, a citation cannot say which version or which authority you are looking at.
Does Vaquill AI offer this through an API?
Vaquill AI serves US primary law through a REST API, an MCP server and SQL: statutes, regulations, constitutions, court rules, executive orders and agency guidance, with a citation and a source link on every section. See the US primary law API.
New legal AI guides, weekly.
Further Reading
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.