Loading content
Microsoft GraphRAG is a Python framework for transforming text corpora into a knowledge graph and executing graph-based retrieval workflows. It supports local search for entity-centric questions and global search for broader summarization. Distinctive elements include its indexing pipeline, community detection, and emphasis on structured context generation.
Use this profile to understand the building block briefly, place it in the model, and open related building blocks.
Technical building block: can be automated, integrated, or operated.
Concrete cog in the system that works inside larger relationships.
Microsoft GraphRAG is a retrieval-augmented generation approach that uses a knowledge graph and community summaries to search documents.
Microsoft Research developed GraphRAG and released it as a research paper and open-source repository in 2024. It addressed how retrieval-augmented generation could answer global questions over large private text corpora, combining entities, relationships, and communities for that purpose.
An indexing pipeline extracts entities and relationships, builds a graph, and summarizes its communities. For a question, GraphRAG selects local context or global community summaries; a language model turns that context into a source-grounded answer.
Documents are transformed into entities, relationships, and summaries.
Tightly connected graph regions are detected and summarized as topic clusters.
Local search handles details, while global search uses community summaries for corpus-level questions.
GraphRAG can help with large, connected document collections where plain vector search lacks a corpus-wide view. Indexing cost, model errors, privacy, and checking generated answers remain important operational responsibilities.
Where this building block is located in the topic model.
No structure path available.
Explore how this building block connects to concepts, methods, technologies, and tools.
These sources establish the term and its professional meaning.
All direct connections of the current building block in a compact text view.
This classification shows where the building block typically matters, how demanding it is, and what kind of impact it has in the model.
The level within the organization (enterprise, domain, team) at which the AssetBlock is applied.