Data licensing

For legal software and legal AI companies. You can license the immigration law library behind Passim: decision texts as files or through an API, and the citation graph and citator through the API.

What you can license

  • API access for your productYour product calls Passim with its own keys, and usage is metered. Keys can be scoped, given an expiry, and revoked one at a time.
  • Bulk text files, refreshed on a scheduleDecision texts and rule-based metadata, delivered as files, with alien registration numbers masked. We’ll agree the format and the schedule with you.
  • The citation graph and ImmCite statusesImmCite is Passim’s citator: what later decisions did to a precedent, drawn from 287,121 citation links between decisions. Both come through the API only, not as files.
  • Model-training licensesTexts only, by written license. Summaries and citator readings written with AI models are available through the API and are not licensed for model training.

What’s in it

1,192,029 decisions, counted November 16, 2026. The Library page lists every source, including statutes, regulations, and agency guidance.

Board of Immigration Appeals and Attorney General precedentsVolumes 1 to 30, 1940 to the current volumeDaily4,533decisions
Unpublished BIA decisions2017 to the present, as EOIR has released themAs EOIR releases them286,926decisions
Administrative Appeals Office decisions2005 to the presentDaily169,907decisions
Federal circuit court opinions1920s to the presentDaily135,474opinions
Federal district court decisionsHistorical and current decisionsWeekly84,487decisions
U.S. Supreme Court opinions1830s to the presentAs decided3,017opinions
OCAHO and BALCA decisionsBALCA from 1987, and OCAHO’s published decisionsDaily14,878decisions
Immigration judge decisionsReviewed setAs reviewed296decisions
State court decisions on criminal statutesSelected state courtsIn batches, by state492,497opinions

Where it comes from and how it’s kept current

The library is built from public legal materials, drawn from more than 120 official sources. We check most of them daily, and the table gives each collection’s schedule.

We also check and repair text and metadata against the preserved originals, and we keep earlier versions so corrections can be checked. Coverage varies by source and year, and some records may be incomplete.

Terms in brief

  • Passim’s terms of service bar extracting the database in bulk, republishing or reselling it, and using it to train a model without a written license. A license is what permits them.
  • Courts, agencies, and other publishers keep whatever rights they have in the underlying materials. Passim keeps the rights to its compilation, metadata, summaries, and research signals.
  • Summaries, classifications, and citator statuses are Passim’s work, some of it produced with AI. They can’t be presented as the statement of a court or agency.
  • Each data license is its own written agreement, separate from the subscription terms.
  • Passim is not yet independently audited. The Security page lists every control and says which are planned rather than in place.

Ask for a quote

There’s no price list. A quote depends on which collections you need, how they’re delivered, and what you’ll use them for. Chris Hammond, the attorney who built Passim, replies within one business day.