Skip to main content
GET
Search and page through every rule we hold, across all courts. Filter by court, judge, rule type, workflow phase, case type, topic, or text. Results are newest first and paged with limit and offset. To keep only the rules that say something about artificial intelligence (AI), send topic=ai_use (see Topics below). To read the requirements of an official rulebook such as the California Rules of Court instead, send rulebook_id (see Rulebook requirements below).
string
Court identifier, such as il-cook-circuit, ca-los-angeles-superior, or edny.
string
Judge, calendar, department, or courtroom slug. Court-level rules use court.
boolean
When judge_slug is supplied, include court-level rules where judge_slug is court. Defaults to true.
string
Filter by structured rule type, such as ElectronicFilingRule, FilingTimingRule, ServiceRule, FilingFeeRule, or CourtesyCopyRule.
string
Filter by workflow phase, such as FILING, CASE_INITIATION, or MOTION_PRACTICE.
string
Filter by case type. General rules and rules without an explicit case-type limit are included.
string
Search summary, source text, tags, and structured data. Every word must appear in the rule. Example: rejected cure.
string
Only rules that have this topic. Today the only value is ai_use: rules in which a court or judge says something about AI in filings and proceedings. Applied before the count and the paging, and it combines with the other filters. Cannot be combined with rulebook_id.
string
Rulebook identifier from /rulebooks, such as california-rules-of-court. Returns the requirements read from that rulebook’s current rules instead of a court’s documents. Cannot be combined with district_id, judge_slug, include_court_rules or topic.
string
With rulebook_id: only the requirements read from this rule, by its official number, such as 3.1113. Requires rulebook_id.
number
Maximum rules per page. Defaults to 100, maximum 500.
number
Number of matching rules to skip. Defaults to 0. Pass meta.next_offset from the previous page to read the next one.

Response fields

array
Validated extracted filing rules with provenance.
object
The validated extracted rule payload. It includes source text, summary, tags, workflow phase, severity, structured data, and citation details when captured.
object
What the rule is about, keyed by topic. Present only on a rule that has a topic, and left out otherwise. topics.ai_use holds kind, label and fields; see AI use rules.
string
Official source URL for the rule.
string
Source document version used for the extraction.
object

Paging

Read a whole court by following next_offset until it is null:
Rules are ordered newest first. If the court’s documents are re-checked while you page, a rule can move between pages; re-read from offset=0 when you need an exact snapshot.

Filtering behavior

When judge_slug is provided, the endpoint returns rules for that unit. With include_court_rules=true, the response also includes court-level rules for the same court, so a client can combine the court’s rules with the rules of a judge, calendar, department or courtroom. The endpoint returns the extracted rules as they are. It does not yet resolve conflicts between rules or work out the final set of rules that applies to a filing.

Rulebook requirements

Send rulebook_id to read the requirements of a rulebook: the filing requirements we read from the text of its rules, with the rule that sets each one. Add rule_number for one rule. The other filters work the same way: logic_type, workflow_phase, case_type, q, limit and offset. See Rulebooks for the rulebook, rule and requirement terms.
What differs from the court results above:
  • rulebook_rule names the source. It holds the rulebook, the rule number, how to cite the rule, a link to the rule’s page, and the revision_id of the version of the rule the requirement was read from. While that version is current, it equals the revision_id of the rule.
  • No court or judge fields. A rulebook’s requirements belong to a rule, so results carry no district_id, judge_slug, source_url or doc_kind.
  • Order. Results are ordered by rule number as text, then by id. The order is stable between calls but is not the order of the rules in the rulebook; get that from /rulebooks/{rulebook_id}/rules.
  • No page numbers. Rule text is not paginated, so content.citation_source never carries a page.
  • Topics. A requirement carries topics when it has them, as a court’s rule does. The topic filter does not apply to requirements.
  • A rule with no requirements returns 200 and an empty rules array. Its extraction_status on the rule says whether the text has been read.
  • Meta echoes rulebook_id and rule_number instead of the court filters.
A request cannot name both a rulebook and a court or judge:
400
rule_number without rulebook_id is also a 400, and so is topic with rulebook_id. An unknown rulebook, or a rule number the rulebook does not have, is a 404.

Topics

A rule that is about a topic carries a topics field, keyed by topic. Today the only topic is ai_use: rules in which a court or judge says something about AI in filings and proceedings. AI use rules explains kind, label and each field, with a full response. Send topic=ai_use to keep only the rules that have the topic:
  • Before the count and the paging. meta.total and the pages are those of the rules that have the topic, and meta.topic echoes the filter.
  • With the other filters. topic combines with district_id, judge_slug, include_court_rules, logic_type, workflow_phase, case_type and q.
  • Every kind of rule. The filter keeps every rule that has the topic, whether it asks something of the filer or not. Check topics.ai_use.kind for the rules that do.
  • No topic, no field. A rule that does not have the topic has no topics field.
A topic the API does not offer is a 400:
400
A request cannot name both a rulebook and a topic. Rulebook requirements carry topics, but they cannot be filtered by topic:
400

Limits

  • Text and case-type searches: q and case_type read every candidate rule before matching, up to 10,000 candidates. The limit covers any single court. A search across all courts with a common word returns 422; add district_id, judge_slug, logic_type, workflow_phase, or more specific words.
  • Unknown court: a district_id we do not map returns 404.
  • Unknown rulebook or rule: a rulebook_id that is not in /rulebooks, or a rule_number the rulebook does not have, returns 404.
  • Unknown topic: a topic the API does not offer returns 400.
For one judge’s complete rule set in a single call, grouped into court-wide and judge layers, use /rules.