Back to Blog
    Perspectives

    Public Case Law Isn't a Moat. Your Firm's Confirmed Facts Are.

    Why CourtListener alone will not beat EvenUp, and what will.

    Bijan Sirleaf
    November 2, 2026
    7 min read
    Public Case Law Isn't a Moat. Your Firm's Confirmed Facts Are.

    Every competitor can search the same public opinions. CourtListener is useful, and Marlowe uses it for live legal research, with citations marked verified only when they match a real returned case. That alone does not create a durable advantage. A public corpus is, by definition, a corpus other people can use too. If the only difference between two tools is that both can search public opinions, there is no moat there, only a shared starting point.

    The data a competitor cannot download sits inside the firm: facts an attorney confirmed, corrected, or disputed on the Facts tab; page-cited details pulled from that firm's own medical PDFs; demand drafts that started from confirmed facts first; and case outcomes the firm chooses to keep, stripped of client identifiers. The honest takeaway is simple. Use public case law for research. Build the moat from confirmed facts and attorney judgment, with the firm's consent.

    Why isn't public case law a moat?

    A moat, in this sense, is something a rival cannot copy by signing up for the same public data. CourtListener opinions are public. Live search against them is a good research method, and it is a better method than inventing citations. It is not a secret.

    Legal research on real court opinions, and its limits, is the product-level description of that workflow. Marlowe searches CourtListener live, analyzes only the cases the search returned, and marks a citation verified only when it matches a real returned case. Those limits are features, not fine print. They also explain why research alone will not set one PI operations tool apart from another vendor that can also query public law.

    This is the idea behind the original subtitle of this post. CourtListener alone will not beat EvenUp, or any other product in the same market, because public case law is not exclusive. The comparison page is the place to judge the actual products, using each vendor's own description and a test on your files. This post is about the data argument underneath that comparison, not a claim that one vendor has already won it.

    What data actually stays inside the firm?

    The work a PI firm does on its own files is not in CourtListener.

    An attorney can confirm, correct, or dispute a proposed fact on the Facts tab. Those decisions are this firm's reading of this record. How attorney corrections should improve the next draft is the supervision loop. Demand drafts are written to prefer confirmed facts first, so the next letter starts from what the attorney already approved.

    The facts themselves are often pulled from the firm's own medical PDFs: treatment, injury, and bill details, with page citations where possible. What Marlowe pulls from medical PDFs today is the live extraction path. Those pages are not a public corpus. They are the client's records, held by the firm, and they should stay under the firm's control.

    If the firm later chooses to keep case outcomes, stripped of client identifiers, that too is firm data, not something a rival can scrape from a public opinion database. Choosing to keep that information is the firm's decision, and it requires the firm's consent. We do not treat client files as a training set by default. Consent is the condition, not an afterthought, because the firm's own facts are the only data in this picture that a rival cannot simply download.

    What does Marlowe do with that data today?

    The live workflow is already built around the firm's file, not around a private copy of the case law.

    Marlowe searches CourtListener live and analyzes only the cases returned. Case prep that stays tied to the file builds a matter timeline and can include a short list of relevant precedents, next to the pulled treatment, injury, and bill details and a list of what is still missing. Attorneys can confirm, correct, or dispute each fact before a demand uses it.

    Current rates for the case preparation unit are published on the Nimbus Marlowe page. Research support and the Facts tab sit inside that operational work. They are not a separate promise that public law has been turned into a unique library.

    This is also how an agentic operations layer is supposed to work. The layer handles repeatable work around the file. The attorney stays responsible for the facts and the legal decisions. Every output stays a draft until that review happens.

    What is not built yet?

    Three limits belong in the same breath as the live workflow, because they are the claims this argument is most often stretched into.

    There is no stored library of every CourtListener case. Live search is not a proprietary case collection, and bulk storage of that data for a paid product would need a commercial arrangement with Free Law Project.

    There is no full medical-chronology product sold as its own finished suite. Case prep produces a draft timeline and page-cited facts. That is not the same thing.

    Automatic learning from attorney edits into a shared model is not built yet. Firm playbooks, which would store rules the firm writes and approves, are still shipping. The Marlowe page says so directly: we do not claim firm-wide learning yet. The moat path is the firm's confirmed facts and judgment, collected with consent. The live product is the review loop, not a model that has already learned those rules automatically.

    How should a firm compare this to other tools?

    If you are choosing software, do not treat public-case-law access as the decisive feature. Other vendors can point at public opinions too. Compare the work around the file: what is extracted, what the attorney reviews, what happens after a correction, and how a demand is drafted from confirmed facts.

    The EvenUp comparison and the guide to AI needs for small PI firms are the places to do that work, using each vendor's own description and a test on your records. This post does not claim that Marlowe has already won that comparison, and it does not repeat vendor marketing numbers. It makes a narrower point. Public case law is useful research. It is not the moat. The moat, if there is one, is the firm's own confirmed facts and the attorney's judgment, used with the firm's consent.

    If you want to see the Facts tab and case preparation on one of your files, start on the Marlowe page. The opinions stay public. The confirmed facts stay yours.

    Frequently asked questions

    Does CourtListener research create a durable advantage?

    It is useful research support, and Marlowe uses it with citations marked verified only when they match a returned case. It does not, by itself, create a durable advantage, because the underlying opinions are public.

    Are firm playbooks or automatic learning from edits live?

    No. Automatic learning from attorney edits into a shared model is not built yet, and firm playbooks are still shipping. We do not claim firm-wide learning.

    Does Marlowe store a private CourtListener library?

    No. There is no stored library of every CourtListener case. The research path is live search, not a proprietary case collection.

    What data does this post say actually compounds?

    Facts an attorney confirmed, corrected, or disputed; page-cited details from the firm's own medical PDFs; demand drafts that started from confirmed facts; and case outcomes the firm chooses to keep, stripped of identifiers, with the firm's consent. The live product is the review loop around those facts, not a finished learning system.

    Should I pick a vendor based on public case law access alone?

    No. Use public case law for research, and compare tools on the file work, the review loop, and the demand draft. The EvenUp comparison and the small-firm AI needs guide are the practical places to do that, with a test on your own records.

    Put your firm's operations on autopilot. Reviewed by you.

    Nimbus Marlowe is the agentic operations layer for personal-injury firms. The CRM is free, pricing is flat and per-unit, and every output is a draft your attorneys review and adopt.

    Contact us

    Tell us what you are trying to automate and we will get back to you.