Recompute stored memory embeddings to match a new embedding model.
npx skills add https://github.com/emmahyde/memesis --skill memesis-reindex-vec
Please help me install this Agent Skill. Skill: memesis reindex --vec Source: https://github.com/emmahyde/memesis/tree/main/skills/reindex Command: npx skills add https://github.com/emmahyde/memesis --skill memesis-reindex-vec
It enables updating all stored memory embeddings to match a new or changed embedding model, ensuring retrieval accuracy and consistency.
Use the reindex --vec skill to refresh memory embeddings after changing your embedding configuration.
High-intent search queries and answers about installing and using this skill.
To re-embed stored memories after an embedding model update, you recalculate all existing memory vectors using the current model to maintain retrieval accuracy and consistency.
Memory retrieval accuracy drops after altering embedding parameters because existing stored vectors no longer match the new model's vector space, requiring a full re-indexing process.
The best way to update a vector database after switching embedding models is performing an atomic table swap, which executes an in-place update to maintain database integrity during re-embedding.
Yes, you need the core.embeddings, core.vec, and core.database dependencies to execute memory re-indexing, as they handle embedding generation, vector storage, and database operations.
You can update memory embeddings in-place without database corruption by using an atomic table swap mechanism, which safely replaces the vector memories table during the re-indexing operation.
dotlab-hqBuild RAG pipelines that ingest, chunk, embed, and retrieve documents for LLMs.
MSamiulHasnatIndex embeddings and perform context-aware similarity queries for document retrieval.
multiversxImplement an in-memory token ledger using ManagedMapEncoded and ManagedVec for MultiversX transactions.
hiyenwongIndex local SKILL.md documents into a SQLite vector search index.
