AI GitHub
Cognee for building memory from documents
Turn inputs into connected knowledge and retrieve that context for AI applications.
Key Takeaways
- Cognee organizes processing around ingestion, a knowledge-building step and search.
- Choose the storage backends and processing options for a disposable corpus.
- Add a few notes with repeated names and one deliberately conflicting claim.
- A generated graph can merge separate entities or preserve outdated relationships.
How it works
Cognee organizes processing around ingestion, a knowledge-building step and search. Its pipeline can store semantic and relational representations of the same data.
What you need
Choose the storage backends and processing options for a disposable corpus. Keep a mapping back to source records so results can be audited after an update.
What to try first
Add a few notes with repeated names and one deliberately conflicting claim. Inspect the returned evidence before and after updating one note.
Put a plausible distractor in the corpus with a similar entity name. Inspect whether the knowledge-building step keeps the two entities and their relationships separate.
Where to look for failures
A generated graph can merge separate entities or preserve outdated relationships. Test update and deletion behavior against the underlying source records.
Related Franklin pages
Use a question that needs evidence from several sources and another that needs an exact quotation. This helps show when building intermediate relationships or summaries is worth the effort.
- Study a research pipeline that generates questions, gathers sources and drafts a structured article. STORM for researching and drafting cited articles
- Index relationships as well as text for questions that span multiple source passages. LightRAG for graph-assisted document retrieval
For the full path from source documents to generated answers, start with Franklin’s glossary. Retrieval-augmented generation
Project documentation
Try it for your workflow
Explore the project’s documentation for setup and usage, or save this page to a collection for later.
Read the project documentation ↗