Vaquill AI vs Midpage: US legal data API comparison

Both Vaquill AI and Midpage sell US primary law as an API. Midpage began in case law and added codified law in July 2026; Vaquill AI began in codified law and added the federal administrative layer. Every Vaquill AI figure comes from GET /us/statutes/coverage (measuredAt 2026-10-12T16:46:07+00:00) or the deployed OpenAPI document; every claim about Midpage links to their own documentation, read on 2026-10-12.

Reviewed and corrected on 12 October 2026. This is general information, not legal advice.

For OpenLaws, see Vaquill AI vs OpenLaws and the three-way comparison.

TL;DR

• Vaquill AI serves 5,454,678 sections across 23 corpus types, indexed as 20,781,207 retrieval passages.
• 96,966 guidance documents across 53 named sources and 26 agencies, plus 100,920 adjudication sections across seven bodies, led by NLRB at 66,597 and FTC Part 3 at 18,023.
• 20,604 state session laws from the 2025 sessions across 50 jurisdictions, and federal local court rules from 52 courts. Midpage lists federal public laws from 2013 and no court rules.
• 105,087 state Attorney General opinions across 43 jurisdictions, state court rules across 56, state insurance bulletins across 49, and Puerto Rico regulations at 111,484, ahead of California at 74,764. Midpage lists state insurance bulletins as upcoming and none of the rest.
• Up to 148 fields per section, including statute-to-regulation joins in both directions, RIN rulemaking chains, and character and page offsets into the source file. Midpage's LawDocument has 21 fields plus a versions array.
• 310 watchable law-change boards with webhook delivery. Midpage documents no alerting.
• Midpage carries 16M+ opinions and 6M+ codified documents, an AI citator and on-demand PACER retrieval. The Vaquill AI data API is scoped to codified and administrative law by design.
• Midpage stores historical versions of statutes and regulations, behind a separate historical entitlement. Vaquill AI serves 31 published GPO annual editions of the CFR and reconstructs statute text.
• Agency guidance overlaps heavily, and Vaquill AI carries guidance families Midpage does not list, including banking supervision, FinCEN, CFTC, export controls and HIPAA. Midpage adds SEC staff guidance, CBP customs rulings, FDA, HUD, OSHA and DOJ ADA guidance.
• Vaquill AI issues a key from its console without a sales call. Midpage publishes a $1,000/mo tier for internal use and prototyping, and licenses production use by contract.

Data coverage

Section counts are distinct provisions. Midpage figures are from its developer, data-quality and practice-area pages and its public coverage pages, read 2026-10-12. The main coverage page at app.midpage.ai/coverage/us sits behind a bot check, so it was read through a text extractor; the federal sub-page is linked.

The two headline totals are not measured on the same definition, so they are not directly comparable. The Vaquill AI figure is what GET /us/statutes/coverage publishes as sectionCount, and that endpoint is free to call and authoritative over any published page.

Both corpora come from official sources. Midpage states that statutes and regulations are collected from GovInfo and official state repositories through 300+ scrapers, with courts visited six times a day. For states whose official code is published only through LexisNexis, LawSites reported that Midpage reconstructs the code from session laws, with Georgia, Tennessee, Mississippi and Arkansas in progress as of 2026-07-23.

Federal agency guidance

Neither corpus is a superset of the other, and the overlap is wide. Vaquill AI entries are the source enum in the deployed OpenAPI document; Midpage entries are from its federal coverage page, read 2026-10-12.

Vaquill AI carries banking supervision, AML, derivatives, export controls, HIPAA, product safety and labor relations, none of which Midpage lists. Midpage adds SEC staff guidance, CBP customs rulings, FDA, HUD, OSHA and DOJ ADA guidance.

The 26 agencies behind those 53 sources, named so you can check whether yours is here: IRS (six sources, including written determinations and the Internal Revenue Manual), DOJ (Justice Manual, Antitrust leniency, business review letters, and the Merger Guidelines with the FTC), FTC (advisory opinions, policy statements, HSR premerger interpretations, advocacy filings), NLRB (General Counsel and Division of Advice memoranda), CFPB (circulars, supervisory guidance), Federal Reserve, OCC, FDIC, FinCEN (six sources, including the beneficial-ownership set), CFTC, SSA, CMS, HHS Office for Civil Rights (HIPAA guidance, FAQs, resolution agreements), EEOC, DOL Wage and Hour, USPTO (MPEP and TMEP), USCIS, US Copyright Office (circulars and the Compendium), BIS, DDTC, OFAC, CPSC, FCC, FERC, DOE, and the DFARS PGI. Vaquill AI additionally carries banking supervision, AML, derivatives, export controls, HIPAA, product safety and labor relations, none of which Midpage lists.

Scope notes

The Federal Register notices corpus is a curated slice, not the whole series: it holds notices that act on a rulemaking and not the Paperwork Reduction Act collections, meeting notices or permit applications that make up most of the series.

The Internal Revenue Manual is agency procedure, not law. The Statutes at Large corpus is law as enacted, carrying goodLawStatus: "unknown" deliberately, and runs 1789 to 2026 across every bound volume GPO publishes, 137 in all. Volumes 7 and 8 carry no session laws at all, and that treaty material is served under the treaty corpus rather than here. Every CFR annual-edition point is superseded law by construction.

The Vaquill AI adjudication corpora carry no verified doctrinal currency: goodLawStatus is unknown on most of them by construction rather than asserted.

Versions and currency

Midpage stores versions and derives currency at request time. Per its Laws API documentation there is no stored is_current flag: a document is current only if it is not marked historical, its effective_date is not in the future, no newer version is already in force, and it is not a repeal tombstone. get returns a versions array carrying each version's id, effective date and publication date; an older version is fetched by id.

Two limits are stated in the same document. Historical access requires an API key with historical entitlement; without it the widening flags and historical id lookups return 403 and versions comes back empty. effective_date is populated only when the source states it, because several state codes publish session-year history instead.

For regulations, Vaquill AI serves GPO's officially published annual editions of the CFR: 31 editions from 1996 to 2026, all 49 titles, as printed rather than reconstructed. asOf=YYYY-MM-DD resolves a date to an edition, sets asOf.engine to stored_edition and returns asOf.source: "published". editionsObserved is an enumeration and never a range, so a date in an unobserved gap returns asOf.isBounded: false rather than the nearest year held.

For statutes, the same parameter reconstructs rather than archives, rebuilding earlier text from the before-side of the first change observed after that date and returning an isBounded flag when the answer rests on when change capture began. The mechanics are in the OpenLaws comparison.

Endpoints and capabilities

Midpage's codified law sits behind three endpoints sharing a base URL and bearer key with its case-law endpoints: /laws/search, /laws/get and /laws/collections. Its current MCP contract is v4, served at the unversioned URL; searchLaws and analyzeLaw keep their v3 contracts inside it, and v3 stays available.

Midpage's analyzeLaw MCP tool requires a question and returns verified verbatim passages rather than raw text; the REST get returns the markup.

GET /us/statutes/section/{actId} returns up to 148 fields, and search takes a fields list so a response carries only the ones you parse. Midpage's OpenAPI LawDocument has 21 fields: id, parent_id, path, node_type, number, title, sort_order, collection_id, collection_name, collection_type, state, country, citation, register_citation, effective_date, publication_date, is_current, is_historical, html, url and source_url, plus a versions array. The groups with no counterpart there are the cross-reference graph (statutoryAuthority, implementingRegulations, crossReferencesUsc, crossReferencesCfr and 6 more), Federal Register rulemaking metadata (12 fields including RIN, docket, significance and corrections), provenance (granuleId, packageId and character and page offsets into the source file), and supersession chains. Search takes 23 filters, including changedSince as a sync primitive and actStatus and excludeRepealed for dead text.

Verifying citations at document scale

POST /us/statutes/resolve takes up to 50 Bluebook citations in one call and returns a verdict on each, duplicates collapsed before pricing. Every input gets an entry, so an unresolved citation is a reported answer rather than a silent omission. That is the shape of the problem when you are checking a brief, a memo or a model's output rather than looking one citation up.

GET /us/statutes/section/{act_id}/body?format=content splits a section into content, sourceCredit and notes using the publisher's own field delimiters, and returns only the operative text. The split matters more than it sounds: on 17 U.S.C. § 107 the statute text and credit line are a small fraction of a body that is mostly committee reports and guidelines printed as notes. Feeding the full body to a model invites it to quote a note as the law in force.

Midpage's documentation, as reviewed, describes no equivalent for either operation.

The MCP server generates its tools from the same OpenAPI document, so a new endpoint becomes a tool without a package upgrade. It also ships resources and prompts, including a cite_check prompt and a good_law_check prompt, which clients that support them get without spending a tool call.

A short history of who may own the law

Before you read any licence, it helps to know where the law draws the line between what is free and what can be sold. The line is older than the internet.

In 1834 the Supreme Court decided Wheaton v. Peters, one of its earliest copyright cases. One of its former reporters of decisions, the official who gathers and publishes the Court's opinions, sued his successor, claiming a copyright in the Justices' opinions. The Court said no. As the Court itself later summarized it, it held unanimously that "no reporter has or can have any copyright in the written opinions delivered by this court."

In 1888 Banks v. Manchester extended the idea to state courts. A judge's work, the Court said there, "constitutes the authentic exposition and interpretation of the law, which, binding every citizen, is free for publication to all." The same term, in Callaghan v. Myers, the Court found a limit. A reporter could hold a copyright in explanatory material the reporter created personally, such as headnotes (short summaries placed above an opinion), because the reporter has no authority to speak with the force of law.

That pair of rulings is the root of legal publishing. The words of the law belong to everyone. What a publisher writes on top can carry its own terms.

The idea came back to the Supreme Court in 2020. Georgia's official code includes annotations under each statute, summaries of cases and opinions that a private publisher drafted under contract with a state commission. Georgia claimed a copyright in them. In Georgia v. Public.Resource.Org, decided April 27, 2020, the Court ruled 5 to 4 against Georgia. Its reasoning: "The animating principle behind this rule is that no one can own the law." Because the commission acted as an arm of the legislature, even the non-binding annotations fell inside that rule.

Two cautions. The ruling concerns who can claim copyright in works made by lawmakers and judges. It does not say that a paid service built on that law must hand you everything it produces, because a service comes with a contract you agree to. And this is a general history, not legal advice for your situation.

What a licence should say before you build on it

Free law does not make every licence simple. A commenter on Hacker News, writing in March 2026 in a thread about a dataset of Spanish legislation, made the business point plainly: compiling legal data takes "a lot of manual work" and "the cost of that work becomes a barrier to entry." That effort is part of why providers attach conditions to their data, so the conditions deserve a careful read.

Before you build a product on any legal data source, get written answers to these:

• Can your customers see the text inside your product, or is use limited to your own staff?
• Does the plan allow production use, or only internal use or prototyping?
• May you cache or store copies, and what happens to them when you cancel?
• May you use the data to train or tune a model, and are there carve-outs?
• Is attribution required, and where must it appear?
• What notice do you get before terms or prices change?

If the licence is silent on one of these, ask. An unanswered question is much cheaper to settle before launch than after.

Access

Coverage discovery, boards and watch management are free to call on the Vaquill AI API.

Data model

Vaquill AI scopes by corpusType (23 tokens) plus a 2-letter state, and an optional source for corpora folding several bodies of law into one token. Each section has an actId that holds through amendment and repeal; a renumbering changes it, and GET /us/statutes/resolve finds the new one by citation.

Midpage's replica is organized into three PostgreSQL schemas: cases (current, adding roughly 17 million docket entries), opinions (deprecated) and laws (collections, documents as an ltree, version relations).

Summary

• Midpage serves case law, an AI citator, a case citation graph and PACER dockets, with 39 federal administrative, military and other tribunals listed on its coverage page.
• Midpage lists SEC staff guidance, CBP customs rulings, FDA, HUD, OSHA and DOJ ADA guidance, which are not among the Vaquill AI sources.
• Vaquill AI serves the following, which are not listed in the Midpage pages reviewed: rulemaking metadata (RIN, docket IDs, statute-to-regulation joins), law change alerts with webhooks, state session laws, state Attorney General opinions across 43 jurisdictions, state court rules, federal local court rules, state insurance bulletins, Puerto Rico and four more territories, SEC Commission opinions, DOJ OLC opinions, FTC Part 3 decisions, Statutes at Large back to 1789, and banking, FinCEN, CFTC, export-control, HIPAA, FCC, FERC, DOE, CPSC, NLRB memoranda, SSA and DFARS PGI guidance.
• Vaquill AI is self-serve, with free coverage and board endpoints. Midpage publishes a $1,000/mo prototyping tier and licenses production use by contract.
• Both serve the US Code, CFR, state statutes, state regulations, constitutions, the Federal Register and a shared subset of federal agency guidance, and both source from official government publishers.

FAQ

Is Midpage a data API or an application?

Both. The application is sold on published subscription tiers. The data business is separate, sells through a REST API, an MCP server and a SQL replica, and is scoped by sales.

Does Midpage cover statutes and regulations?

Yes, added in July 2026 for federal plus all 50 states and DC, covering statutes, regulations, constitutions and agency guidance; Midpage lists Tennessee, Arkansas, Georgia and Mississippi statutes as upcoming. The REST laws endpoints share a base URL and bearer key with the case-law endpoints, laws search is BM25 keyword, and the laws tools sit inside its v4 MCP contract.

Which one has point-in-time law?

Both, by different mechanisms. Midpage stores historical versions of statutes and regulations where the source has them, gated behind a separate historical entitlement. Vaquill AI serves 31 published GPO annual editions of the CFR, 1996 to 2026, for regulations, and reconstructs statute text with a bounded flag.

Does Vaquill AI have case law or a citator?

No. The Vaquill AI data API is scoped to codified and administrative law, and retrieves by citation.

How do the two access models differ?

Vaquill AI issues a key from its console without a sales call and meters per call. Midpage publishes a $1,000/mo tier, rate limited and for internal use and prototyping only, and moves production use onto a custom annual licence.

How fresh is each corpus?

Midpage states that all its law collections are updated weekly, court websites are visited six times a day, and new opinions appear the same day they are published. Vaquill AI reports currency per section in currentThrough, and per source on GET /boards, where each of 310 boards carries its cadence and lastRetrievedAt. The US Code follows OLRC release points, so 42 U.S.C. § 1983 reads current through 2026-09-18 at this update.

Does either support MCP?

Both. Vaquill AI publishes a tool for every endpoint on a hosted server, reachable through an OAuth connector for Claude and ChatGPT with no key to paste, with an API key, or self-hosted through uvx vaquill-mcp. The tools are generated from the OpenAPI document at startup, so a new endpoint becomes a tool without a package upgrade, and the server adds MCP resources and four prompts including cite_check and good_law_check. Midpage serves a v4 contract of five tools, covering case law, dockets, trial filings and laws, keeps v3 available, and supports a shared API key or per-user OAuth.

Does either have state session laws?

Vaquill AI serves 20,604 acts from the 2025 sessions across 50 jurisdictions, searchable by their enacted text, with approval and effective dates, veto outcome, and in eleven jurisdictions the code sections each act changed. Midpage lists federal public laws from 2013 and no state session laws.

Does either have state Attorney General opinions?

Vaquill AI carries 105,087 formal opinions across 43 jurisdictions under corpusType=STATE_AG_OPINION, led by Florida at 20,960 and Alabama at 11,930. Not verified in our October 2026 review whether Midpage lists any state Attorney General opinions. Check its coverage page.

Where does Midpage's data come from?

Official court and government sources through 300+ custom scrapers, with statutes and regulations from GovInfo and official state repositories.

Get started

See the API overview and quickstart, then create a key.

Sources, all read 2026-10-12: Midpage home, developer data, data quality, pricing, CourtListener comparison post, practice pages for labor and employment, immigration and insurance, docs introduction, Cases API, Laws API and MCP tools, plus its federal coverage page and the main coverage page at app.midpage.ai/coverage/us; LawSites, 2026-07-23; Vaquill AI figures from GET /us/statutes/coverage and the deployed OpenAPI document.

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.