HazyResearch/deepdive

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

1979 stars · 533 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

DeepDive is a Stanford-developed system for extracting structured data from unstructured sources and building knowledge bases using distant supervision and probabilistic inference. It lets users declare extraction and inference rules over databases rather than writing feature engineering code.

Use cases

• build a knowledge base from scientific papers
• extract entities and relations from text documents
• populate a database from unstructured text
• train relation extractors without labeled data
• run probabilistic inference over extracted facts
• perform information extraction over dark data

When to choose

• you need to construct structured knowledge bases from large unstructured corpora
• you want distant-supervision-based extraction without hand-labeling training data
• you work with domain-specific text like scientific or legal documents

When to avoid

• you need actively maintained software with new features or guaranteed support
• you want a modern weak-supervision pipeline - consider the Snorkel project instead
• you need a lightweight library rather than a full system with database dependencies

Facets

application · maturity maintenance

machine-learning nlp database etl data-science data-science databases big-data cli information-extraction knowledge-base-construction distant-supervision probabilistic-inference stanford dark-data natural-language-processing linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="HazyResearch/deepdive")