Semantica · treatment network
Citator
How later courts have treated each authority across 2,000 canonical citations. The doubt filter is pre-applied — 105 authorities carry overruling / doubting / not-followed treatments somewhere in the corpus. Clear it to browse everything; expand any row for the citing cases.
No authorities match.
How these claims are derived & verified (methodology)
Source: 20,851 production briefs (19,942 Court of Appeal + 909 Supreme Court), each extracted from the official Kenya Law judgment by GLM-5.3-Flash through the verified pipeline — every quote in a brief is char-for-char matched to the source text (corpus mean grounding 97.7%, median 100%) before publishing.
Evidence chain: every authority count below is built from the casesCited and citationTreatment fields of those briefs. Each count therefore traces to concrete cases — the citing-case lists on this page are the citations; open any one to see its own verified brief and source URL. Nothing here is asserted by the model alone; it is aggregated from verifiable extraction.
Computation: citation graph (222k edges) → centrality (PageRank) and treatment tallies, computed locally. Authority names are normalized by rule + an LLM-assisted pass over the top-cited names (a display-level merge; residual variants can still co-exist).
Not legal advice. Citator status is a data signal from the corpus, not a determination that a case is or is not good law. Verify against the original judgment before relying on it.
A note on accuracy. These pages are generated automatically from our AI-written case briefs and have not been checked by a lawyer. Case titles and citations come from Kenya Law; the counts, topics, propositions and “treatment” labels are derived from the briefs and can be incomplete or wrong. Use them as a research aid and always read the original judgment before relying on anything here. This is not legal advice.