Natural-language query
The user does not need to know an ECLI, a case number, or precise legal terminology.
Find relevant case law for a real legal problem
Lukiora Lex searches relevant court decisions in the case-law catalogue of the Constitutional Court. It helps you quickly find verified sources and the best matching passages.
Lukiora Lex is a specialised case-law search engine. It does not provide legal advice — it returns relevant decisions and their text passages from the current catalogue.
Case-law search
Describe the legal problem as it would be formulated in a legal research note.
For example: When does a flat owner become entitled to compensation for use of the flat after the tenancy ends?
Lukiora Lex is a technology tool for searching court decisions. It does not provide legal advice or legal opinions.
Hybrid searchVerified sourcesRelevant passagesConstitutional CourtSupreme Court
What Lukiora Lex does
Legal research grounded in verifiable documents
Hybrid search
The system combines semantic search with lexical search for legal terms.
Verified provenance
Each result is linked to a specific court decision, identifier and source.
Relevant text passages
The user sees the part of the decision that made the document relevant.
How search works
From a legal query to relevant decisions
Search process
Six connected steps from entering a query to a verified list of decisions.
- 01
Step 1 of 6
Legal query
- 02
Step 2 of 6
Query analysis
- 03
Step 3 of 6
Hybrid search
- 04
Step 4 of 6
Result grouping
- 05
Step 5 of 6
Source verification
- 06
Step 6 of 6
List of relevant decisions
- 01
Step 1
Legal query
02Step 2
Query analysis
03Step 3
Hybrid search
- 04
Step 4
Result grouping
05Step 5
Source verification
06Step 6
List of relevant decisions
Product development
Current status and further development
Lukiora Lex is being gradually extended with further courts and advanced functions for searching Czech case law.
Today — current version
- searches relevant decisions of the Constitutional Court
- combines semantic and textual search
- ranks results by relevance
- shows relevant passages from the retrieved decisions
- lets you open the detail view and the full decision text
- works only with verified decisions from the indexed corpus
Target version
- retrieval of relevant unique decisions in the indexed corpus
- grouping of results by individual decision
- deduplication
- relevance scoring at the level of the whole decision
- dynamic display of results by relevance
- the ability to refine search with follow-up questions
- a transparent explanation of why a particular decision is relevant
Planned
Further case-law coverage
The next aim is to gradually extend search to decisions of further Czech courts. These courts are not yet part of the current search.
- High Court in Prague
- High Court in Olomouc
- Supreme Administrative Court
Technology
Architecture overview for verifiable search
Lukiora Lex separates the presentation interface, the retrieval layer, document provenance and measurable evaluation of results.
Frontend
Next.js, React, TypeScript, Tailwind CSS
Backend
FastAPI, Python
Retrieval
BGE-M3, BM25, RRF, ColBERT, CrossEncoder
Storage
Qdrant
Observability
Prometheus, Grafana
Evaluation
reproducible retrieval benchmarks, provenance validation, regression tests
Who Lukiora Lex is for
A tool for working with case law in a professional setting
law firms
in-house legal teams
academic research
legal-tech developers
case-law analysts
Presentation MVP