$ loading_
Give AI systems local-first long-term memory with text and embedding retrieval.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "NarratorDB" yet — see the docs or source repo.
Use NarratorDB to retrieve past memories related to 'user preferences and previously discussed feature requests', and return the top 5 most relevant results.
A relevance-ranked list of past memories for the agent to continue the conversation or make decisions.
Help me design a local long-term memory setup with NarratorDB: use SQLite FTS5/BM25 for retrieval and explain when to enable SentenceTransformer embeddings.
A memory architecture recommendation for AI apps, including when to use keyword retrieval versus embedding retrieval.
Based on NarratorDB's capabilities, compare SQLite FTS5/BM25 retrieval with optional embeddings retrieval, including differences, pros, cons, and use cases.
A clear comparison that helps choose the right memory retrieval approach.
Developers building a local-first AI assistant can use it to store and retrieve historical information so the assistant keeps context across conversations. It fits scenarios that want less reliance on remote memory services.
When an AI system needs to quickly find past records by keywords, it can retrieve relevant memories through SQLite FTS5/BM25. This is suitable for applications needing lightweight and interpretable retrieval.
Researchers or developers can optionally enable SentenceTransformer embeddings in addition to keyword retrieval to support more semantic memory recall. This works well when user queries do not exactly match the original records.
It is a local-first long-term memory database for AI systems. Based on the provided description, it supports retrieval with SQLite FTS5/BM25 and can optionally use SentenceTransformer embeddings through an MCP server.
From the description, it relies on SQLite FTS5/BM25 for retrieval; if you want semantic retrieval, it also uses SentenceTransformer embeddings. For exact installation and runtime requirements, see the source repository.
It explicitly provides SQLite FTS5/BM25 full-text retrieval and treats embeddings as an optional capability rather than the only method. That means it fits both keyword-based retrieval needs and scenarios that benefit from semantic recall.
Give AI agents local-first memory with searchable, deterministic recall.
Give AI assistants local-first graph memory with semantic search tools.
Let AI agents keep persistent, searchable memory with editable history across sessions.
Provides persistent local memory for AI assistants with note storage and retrieval.
Give AI coding agents local-first persistent memory management and recall.
Give AI coding assistants durable memory for better context and development continuity.