๐๐๐ฆ๐ ๐๐ผ๐๐ป๐ฑ๐ฎ๐๐ถ๐ผ๐ป ๐ฝ๐๐ฏ๐น๐ถ๐๐ต๐ฒ๐ ๐๐๐ โ ๐ฎ๐ป ๐ข๐ฅ๐ ๐ณ๐ผ๐ฟ ๐๐ผ๐ป๐๐ฒ๐ ๐ ๐๐ป๐ด๐ถ๐ป๐ฒ๐ฒ๐ฟ๐ถ๐ป๐ด ๐ถ๐ป ๐๐๐ ๐ฆ๐๐๐๐ฒ๐บ๐
Today DDSE (Decision-Driven Software Engineering) Foundation releasing CEF (Context Engineering Framework) as an open-source project under the DDSE Foundation.
CEF is our attempt to give Java teams a first-class โ๐ข๐ฅ๐ ๐ณ๐ผ๐ฟ ๐ธ๐ป๐ผ๐๐น๐ฒ๐ฑ๐ด๐ฒโ โ a way to model nodes, edges, relations and chunks as domain entities, persist them into graph + vector stores, and retrieve context for LLMs through graph-aware, hybrid RAG instead of blind top-k similarity.
Just as Hibernate abstracts databases for transactions, CEF abstracts knowledge stores for Context Engineering. Build, test, and benchmark intelligent context models in minutes, without the complexity of enterprise graph infrastructure.
๐ช๐ถ๐๐ต ๐๐๐ ๐๐ผ๐ ๐ฐ๐ฎ๐ป:
Model your domain knowledge graph using simple Java types (Node, Edge, RelationType, Chunk).
Store and query through pluggable backends (DuckDB VSS, Postgres+pgvector, Neo4j and others).
Run GraphRAG-style retrieval that combines graph traversal, vector search and BM25 fallback for robust โnever emptyโ context.
Keep everything Java + Spring Boot friendly for existing enterprise stacks.
This is a research-grade beta (0.5), not a finished platform. Weโre publishing it now because we believe context engineering needs shared, inspectable infrastructure, not just proprietary pipelines and marketing decks.
๐ Project page: https://ddse-foundation.github.io/cef/
๐ GitHub repo: https://github.com/ddse-foundation/cef
If youโre working on GraphRAG, knowledge bases, or agentic systems in Java, Iโd love your feedback, issues, and critiques โ and, if itโs useful, your stars and forks.