What it is
UniCourt is a court-data company in Tustin, California, founded in 2014 and run by co-founder and CEO Josh Blandi. It collects dockets and documents from state and federal courts, including PACER, and links each record to real-world entities: attorneys (from bar data), companies (from Secretary of State filings), law firms, and judges. It then sells that structured record in three ways.
- Enterprise API. UniCourt has sold API access since 2017 and relaunched it as the Enterprise API in March 2023. It covers case search, docket tracking, bulk feeds, and analytics endpoints for attorney, law firm, party and judge data. This is the core product for legal ops and data teams.
- DART (Docket Analytics, Research and Tracking). This is the web app for lawyers, launched in August 2025. In February 2026 UniCourt added judgment research, and in March 2026 it added judgment analytics (damages totals, median awards, time to judgment) and attorney comparison. Both are live only for a selection of state trial courts so far.
- AI access. On 18 September 2026 UniCourt launched a beta plugin that connects its litigation database to ChatGPT Enterprise. UniCourt’s blog describes it as an MCP plugin. It requires an active UniCourt enterprise subscription.
In the September announcement, UniCourt claims more than 5,000 state and federal courts in 46 states and more than 2 billion dockets and documents. In June 2026 it said Firm Prospects would build UniCourt’s litigation history into its attorney and law-firm profiles, which shows the data is resold as well as used directly.
Why it shows up in Legal Ops stacks
- The job is a pipeline, not a search box. An in-house team can use it to watch for new suits against the company or its vendors, to pull outside counsel’s case history into a panel review, and to feed litigation counts into a BI dashboard. All of that needs a structured API feed of dockets with normalized names, which is what UniCourt sells. A research seat cannot do it.
- Entity normalization saves the cleanup. Raw dockets spell the same company five ways. UniCourt says it resolves parties, attorneys and judges to one record each. That is the step teams otherwise do by hand in a spreadsheet before any count means anything.
- It now reaches the AI tools lawyers already use. With the ChatGPT Enterprise plugin, a lawyer can ask about a judge’s or opposing counsel’s docket in plain language without exporting anything. Trellis and CourtListener shipped similar connectors in 2026, so this is now the normal way to buy court data, not a differentiator.
Pricing reality
UniCourt’s pricing page lists three self-serve plans: Personal at $59 a month, Professional at $199 and Premium at $399. All are month-to-month with no long-term commitment. Paying yearly gives two months free, which works out to about $49, $166 and $333 a month. There is no free trial. The plans differ in search allowance, number of users and document access. Pulling a PACER document can carry a fee on top, so ask how court charges pass through before you assume the seat price is the whole cost.
Most legal ops buyers end up on the Enterprise plan, because the API, bulk data and the ChatGPT plugin all sit there. It is quote-only. Several software directories list a starting price of $2,250 a month, but that figure is old and UniCourt does not publish it. Treat it as a floor for planning, not a quote. If that floor still holds, an in-house team pulling data into dashboards should budget at least $27,000 a year, scoped by courts, record volume and API calls.
Best for
Legal ops and litigation data leads at corporate legal departments and law firms who need state and federal docket data, with attorneys, parties and judges matched to clean records, delivered into their own systems by API. The clearest fit is a team building a litigation-exposure dashboard or an outside-counsel scorecard that needs data from many states rather than one.
Versus the alternatives
Bloomberg Law and Lexis CourtLink are the incumbents with the largest docket collections and the biggest share of law-firm research budgets. Pick one of them when lawyers need dockets next to case law, news and company research in one seat, and you do not need a data feed. The fastest-growing entrant is Clio Docket, the former Docket Alarm, which came to Clio with the vLex acquisition and went into Clio Work in August 2026. Pick it if your firm already runs on Clio.
Trellis is the closest head-to-head. Pick Trellis when the job is state trial court research and judge analytics for litigators in one or two states, with published prices and a Claude connector. Pick UniCourt when you need federal and state data together, by API, at volume. CourtListener is free, with an API and MCP server. Pick it for federal opinions and the PACER documents others have bought through RECAP when budget is zero. PacerPro is for capturing your own matters’ filings into the DMS, not for researching other parties. Lex Machina wins on curated federal analytics for motion strategy.
If none fit, start with CourtListener’s free API for federal coverage and a saved PACER search for your company’s name, and log what is missing for a quarter. That list of missing state courts is the spec you take to UniCourt or Trellis.
Watch-outs
- Coverage claims move between announcements. UniCourt’s June 2026 release said 4,000 courts across more than 40 states, and the September release says 5,000+ courts in 46 states. Judgment analytics cover only selected state trial courts. Guard: send your top 15 venues by case volume, get coverage and update frequency confirmed per court in writing, and spot-check 20 recent dockets against the court’s own site during the trial.
- The price that matters is unpublished. Self-serve plans are cheap, but the API, bulk data and ChatGPT plugin all need an Enterprise contract. Guard: get the per-call or per-record rate, the overage terms, how PACER fees pass through and a renewal cap in the order form before you build anything on the feed.
- The ChatGPT plugin is a beta and only works in ChatGPT Enterprise. Guard: confirm with both vendors what query data leaves your workspace and whether UniCourt logs prompts. Pilot the plugin with a small group before announcing it firm-wide, and keep the API as the system of record.
- Normalization can be wrong. Matching “Acme Inc.” to the right Acme, or an attorney who changed firms, is exactly where automated entity resolution fails. Guard: before any count reaches a dashboard, audit a sample of 50 matched records against the source dockets, and flag low-confidence matches for review.
- Court data about people is regulated when used for decisions about them. Using litigation or criminal records to screen job applicants or tenants falls under the US Fair Credit Reporting Act. Guard: keep UniCourt for litigation and counterparty research, and route any people screening through an FCRA-compliant background-check provider.