Legal Document Metadata Extraction for Indexing
Extracts structured metadata fields from a legal document to enable retrieval filtering, with confidence levels separating what was found from what was inferred.
Prompt
Extract indexing metadata from the following document. Wrong metadata is worse than missing metadata — it filters out correct results.
Document: <document> [Paste the document or its opening pages] </document>
Return JSON only: { "doc_type": "contract | judgment | statute | regulations | opinion | tender | other", "title": "", "parties": [{"name": "", "role": ""}], "dates": {"signed": "", "effective": "", "expiry": "", "published": ""}, "governing_law": "", "jurisdiction": "", "language": "he | en | mixed", "clause_count": 0, "has_schedules": false, "key_topics": [], "confidence": {"field": "found | inferred | missing"} }
Binding rules:
- found = the value appears explicitly in the text. inferred = derived from context. missing = not present.
- Never fill a field by guessing. missing is a valid and preferred value.
- Dates in ISO 8601. If the document gives a partial date (year only), return what is known and mark it inferred.
- Party names exactly as written, including the corporate form.
- Do not summarise the document's content. Extraction only.
Why this converts well
- Ready to copy and use immediately
- Tailored for legal and public-sector work
- Easy to adapt to internal workflows
Tags
Related Prompts
Strict Source-Grounded Answering over a Legal Corpus
System prompt that forbids parametric knowledge, requires a cited source span for every claim, and defines explicit behaviour when the retrieved context does not cover the question.
Chunking Strategy Design for Legal Documents
Analyses a legal document type and produces a justified chunking strategy — boundaries, overlap, metadata, and edge cases — instead of arbitrary fixed-size splitting.
Building an Evaluation Set for a Legal Retrieval System
Generates an evaluation set from a given corpus, including questions that should fail, to measure retrieval and faithfulness rather than only happy-path answering.