Remembrances FAQs

Question 1

How does Remembrances handle embedding generation, especially for privacy?

Accepted Answer

Remembrances supports local, privacy-first embedding generation using GGUF models (with GPU acceleration), Ollama, or remote generation via the OpenAI API. Local GGUF models ensure no data leaves your environment, providing cost-effective and secure embeddings.

Question 2

What kind of memory layers does Remembrances support?

Accepted Answer

It provides flexible memory layers including key-value storage, vector/RAG (Retrieval Augmented Generation) for semantic search, and a graph database for complex relationships, all powered by SurrealDB. It also manages knowledge bases using Markdown files.

Question 3

What is Remembrances and how does it help AI agents?

Accepted Answer

Remembrances is an MCP server that equips AI agents with multi-layered long-term memory capabilities. It supports key-value, vector/RAG, and graph memory models, allowing agents to store, retrieve, and process information efficiently.

Question 4

Does Remembrances offer documentation or help for its tools?

Accepted Answer

Yes, Remembrances includes an intelligent help system that provides on-demand documentation for its tools. This significantly reduces initial context token consumption for AI agents while ensuring comprehensive usage guidance is always available.

Question 5

What is the Code Indexing System and its benefits for developers?

Accepted Answer

The Code Indexing System utilizes Tree-sitter for multi-language AST parsing, enabling AI agents to semantically search code, find definitions/references, analyze call hierarchies, and even manipulate code across 14+ programming languages. This empowers AI assistants to understand and interact with codebases like never before.

Remembrances

Remembrances

主な機能

ユースケース

主な機能

ユースケース