Skip to content

Conversation

@phamgialinhlx
Copy link
Contributor

@phamgialinhlx phamgialinhlx commented Jan 2, 2026

Summary

  • Add support for user-provided entities in the retain endpoint that get merged with auto-extracted entities
  • Entities can be passed with each memory item as [{"text": "...", "type": "..."}]
  • User entities are deduplicated (case-insensitive) against LLM-extracted entities before merging

Changes

  • API: Added EntityInput model and entities field to MemoryItem
  • Engine: Updated entity processing to merge user entities with extracted ones
  • Clients: Updated Python, TypeScript, and control plane clients with entity support

Copy link
Collaborator

@nicoloboschi nicoloboschi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for your contribution!

Could you elaborate on the use case of this feature?
While I understand how it works, its important that is clear why and when to use it

@phamgialinhlx
Copy link
Contributor Author

Thanks for your contribution!

Could you elaborate on the use case of this feature? While I understand how it works, its important that is clear why and when to use it

In some cases I already know which entities are in the chunks and want to make sure they get included. For example, critical entities like key product names that I've already extracted using regex - I want to pass those directly rather than hoping the LLM picks them up

@nicoloboschi nicoloboschi merged commit dd59bc8 into vectorize-io:main Jan 5, 2026
16 of 23 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants