LexLint is built by UnGovr, a nonprofit transparency platform that documents governments across 200+ countries. Working at that scale means crawling public sites, processing recordings and publishing under many legal regimes at once, so UnGovr had to know, continuously, which AI, scraping, privacy, cybersecurity, communications, content-moderation, product-liability, digital-accessibility, digital-identity, cloud-service, export-control, age-gating and reuse law applied where. LexLint is the tool it built to lint its own work: it checks what an app declares against all thirteen, and cites the instrument behind every finding. Now any coding agent can use it.
Co-Founder and CEO, UnGovr
Sean built the law library for UnGovr's own needs: a researched, cited record of the law that governs a platform crawling and publishing government records in 200+ countries. LexLint is that library turned into a tool any developer's agent can call. He spent 20 years at Akamai helping build its CDN and bot protections, and is now retired from that side of the industry.
Chief Security Officer, UnGovr
Neil ran security programs at Meta, Capital One, NYSE Euronext, HP and Intel over two decades. In LexLint's law library and tooling he saw what security already has: a CVE database, and the vulnerability scanners that read it. Law that reaches software can be cataloged, versioned and scanned for in the same way, and that is the product he and Sean set out to build.
The rest of the team is on the UnGovr people page. They are engineers and operators rather than lawyers, which is worth knowing before you read a finding.
LexLint is forming an advisory board. Its initial group already includes industry leaders from technology, law and cybersecurity, because every finding has a reader in each of the three.
Before the end of 2026 the group becomes a formal advisory board, and its members will be named on this page. Their advice shapes what LexLint becomes; it does not make a finding legal advice.
A lint that gives legal information, not a legal assessment. Findings are research summaries of public legal sources, with citations and as-of dates. They are not legal advice, they create no attorney-client relationship, and they are never authorization to access any system. The passing state is "no basic issues found", never "compliant". LexLint is in beta. Counsel should start at LexLint for legal teams, and the words this site uses are defined in the glossary, in United States legal usage first.
The legal news wire is swept every day across every jurisdiction in the law library, and a story that outruns what we hold opens a research request rather than writing a law row by itself. The law library is rebuilt and republished every day. Every law carries the date it was last checked against its sources, and the footer of every page gives the range, oldest check included, because a heartbeat that only ever shows its best number is not a heartbeat. How LexLint decides what counts as law, and what sits around it, is on the methodology page.
LexLint does not need access to your code. Your own agent reads your code where it lives and writes the declaration, and that declaration is the whole input: the lint is deterministic, a rule-based match against a researched law library, never a model reasoning about the law. When UnGovr reads a company's public website for it instead, that reading is AI-assisted and still sees no code.
AI does help build what the lint reads. The law library is researched with AI assistance, and citations are verified against primary sources. The pages on this site, including the law notes, are drafted with AI assistance under the direction of UnGovr staff and checked against primary sources. Where a note has had a substantive editorial review, the note says so and gives the date. UnGovr holds editorial responsibility for every page; write to hello@ungovr.org. Illustrations on the Scraping Brief's long reads are generated with AI and carry that claim in their file metadata. The News page carries publishers' own headlines: AI selects the stories and resolves their jurisdictions, it does not write them. This section is the disclosure several jurisdictions in the law library now require of AI-assisted publishing, and a lint should pass its own checks. How every UnGovr and LexLint site is made is on How this site is made.
LexLint is an UnGovr service: the UnGovr terms of service and privacy policy apply to all of it, including the MCP server and the UnGovr key a lint runs on. That key passes through to the UnGovr Open Data API on each call and LexLint never writes it down; the API keeps a request log of its own, set out at https://www.ungovr.org/open-data/api-keys, holding which collection was asked for and a one-way hash of the key, never which jurisdiction you looked up. Apart from the records listed below, LexLint stores nothing you have not deliberately sent it.
/lexlint:lexlint-feedback, which you run yourself when you have something to
tell us, shows you the exact text first and sends only what you approve, recorded
against your account so we can write back. A lint run reaches the portal only at
the close of the run that produced it, after your agent has shown you exactly what
it contains and you have approved the upload, and it is kept against your UnGovr
account, never against a session or a repository.
LEM is a fictional robot, the character in the illustrations on the Scraping Brief's long reads. LEM stands in for the UnGovr crawler, which is real, and every situation LEM is drawn into is one the crawler has met: the sign on the lawn, the padlock on the gate, the contract that will not stop unrolling. There are more pictures of LEM on ungovr.org.