Q2 2026: From Platform to Infrastructure
Moonlit opened Connect in Q2 2026, and within weeks Eudia, Saga and Legal Mind were building on its data, backed by a faster platform and more trustworthy answers. The quarter also closed out Moonlit's first full year, with ARR up more than 10x and coverage now at 110+ sources.
5

Moonlit's second quarter came down to one shift: turning from not only a platform people log into every day into infrastructure other legal products build on top of.
Connect, which we opened on 26 May, is the clearest proof of that shift. But it didn't happen on its own. The platform had to get faster first with a continual growth in high quality coverage.
Customers started building on Moonlit within weeks
Connect runs on the same Data Layer that's always powered Moonlit: same jurisdictions, same reference graph. It ships two ways, an API for engineering teams and an MCP server for legal teams working in Claude, Microsoft Copilot, ChatGPT and other AI assistants.
The API and the MCP server are the mechanism. Who builds on them is the real story here.
What matters more is who decided, within weeks of launch, that building on Moonlit's data made more sense than building their own version of it from scratch. That's a much harder thing to earn than a well-built endpoint, and it's the real test of whether Connect is working.
Eudia builds Expert Digital Twins for enterprise legal and business teams. Before Connect, Eudia's coverage was US-focused. Building out European legal data usually means a dedicated data team and a multi-year runway to get there. Eudia got there in days:
"The Moonlit Data API gave us a complete European corpus in days, expanding Eudia's legal source coverage across the Atlantic." (Siddharth Saxena, Product Manager, Eudia)
Put simply, that's the pitch for the Data API: it buys back years you'd otherwise spend building the thing yourself.
The MCP server solves a related problem for a different kind of user: legal teams doing research directly inside tools they already work in every day. Lennard van Mierlo at Quantoz Payments has been running his research this way:
"I formulate a legal question and Claude works out how to break it down, which sources are relevant and reads the full documents. Not summaries, but the actual text. The citations are always correct. I click through to verify them and they have never been wrong." (Lennard van Mierlo, Legal, Quantoz Payments B.V.)
Saga, our longest-running integration partner, says Connect changed what they're able to build. Legal Mind went live on Moonlit data this quarter too, weaving our coverage directly into their own platform.
Lightspeed had to ship before Connect could
An API is only as good as the platform behind it, and Workspaces and Monitors needed to get faster before we could put a public interface in front of them. April's Lightspeed release was a ground-up rebuild that touched both products.
Monitors v2 came out of it with better change detection, so a filing or amendment now shows up in a client's feed closer to when it actually happens, rather than waiting for the next scheduled sync. A unified activity feed replaced what used to be several separate views. And loading got fast enough that most Workspace and Monitor views now render in under a second, even pulling from a dataset of more than 19 million documents.
That backend work never made it into Connect's own changelog, but it's why Connect could launch a month later on top of a platform that was already fast.
Trust matters as much as speed
Speed gets people through the door. What keeps them coming back for something as consequential as legal research is whether they trust what the tool tells them.
That's most of what May and June went toward.
May added numbered citations to every claim inside Workspaces and Monitors, so clicking one lights up the exact passage in the source document. June took Luna 3 out of beta, with more tool calls per research run, visible thinking steps and the ability to pull an entire research session, complete with every source behind an answer, into a Workspace in one click.
This kind of verification matters more this quarter than it would have a year ago. Regulators and courts across Europe are paying closer attention to how AI-generated content gets sourced. For anyone shipping AI-generated legal content, tracing every claim back to where it came from is turning into a legal requirement.
Fifteen new sources, plus a first step outside Europe
June brought fifteen new sources online across Search, Monitors, Workspaces and Luna, among them the European Central Bank, France's Journal Officiel, Norway's Lovdata, UK Parliament Bills and regulators and courts across Belgium, Spain, Denmark and Estonia.
Fourteen of those extend coverage we already had across Europe. That's the expected direction of travel.
The fifteenth is the interesting one: Brazil's Superior Court of Justice, the first source in Moonlit's coverage from outside Europe. It's one line among fifteen, but it's the one that says something about where coverage goes next.
If you want the detail on how any of this actually gets built, the nightly refresh, source selection, backfill and the enrichment work every document goes through before it's usable, we wrote that up separately: Inside Moonlit's data pipeline.
As of publication, our total coverage stands at 110+ sources.
Q2 closed out Moonlit's first full year
ARR is up more than 10x. Net revenue retention sits at 122%. Churn is at zero, with new inbound now closing from both the US and Asia. Four government clients and two of the Big Four are already running on Moonlit data.
The team grew with it. Bas Geelen joined as Product Marketing Manager, and Giedrius Mirklys moved onto the data team full time after finishing his AI Master's at Radboud. In between, we showed up at Legal Geek Amsterdam, FutureLaw Tallinn and LegalTechTalk London, three rooms full of people all asking some version of the same question: whose legal data can you actually trust at scale.
What this sets up for Q3
That shift is what Q2 actually was. Eudia, Saga and Legal Mind went further than using Moonlit. They built on top of it, and trust was what made that possible: legal teams needed to see and verify every claim before they'd rely on Moonlit for real research. Speed and coverage mattered too, but only once that bar was cleared.
Q3 is about doing this again, at a bigger scale. More companies building on Moonlit's data, more integrations like Eudia's and coverage that keeps stretching past Europe.
More on Moonlit Connect: moonlit.ai/connect

Article written by
Khushi Gadhok
