BetterDB Integrates Valkey For Claude Memory
This weekend BetterDB released an open-source memory layer for Claude Code that uses Valkey for vector search and Bun for runtime. The system hooks into Claude Code lifecycle events, summarizes sessions, generates embeddings, and stores data in Valkey using FT.SEARCH with HNSW, enabling semantic retrieval and automatic resurfacing of relevant memories. It supports self-hosting with Ollama or external LLM providers and applies aging and clustering to limit growth.
Key Points
- 1Implements vectorized memory using Valkey FT.SEARCH with HNSW for semantic similarity retrieval
- 2Summarizes sessions and generates embeddings to avoid repeated context and capture essential history
- 3Provides a self-hostable Bun-based memory layer, Ollama-compatible, with aging, clustering, and merged memories
Scoring Rationale
Provides a practical, self-hostable memory layer with direct usability; limited novelty and single-source release reduce broader impact.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
