Lake of Embeddings FAQs

Question 1

Does Lake of Embeddings require an internet connection or API keys to function?

Accepted Answer

No, not necessarily. Lake of Embeddings is offline-first by default, generating embeddings locally using `sentence-transformers` without needing an internet connection or API keys. Optional support for OpenAI embeddings is available if you prefer, which would require an API key and internet access.

Question 2

What is Lake of Embeddings?

Accepted Answer

Lake of Embeddings is a powerful tool that enables local semantic search over your personal knowledge bases. It indexes your notes and databases into a queryable embedding store, allowing you to find content by meaning rather than just keywords.

Question 3

What types of personal knowledge can Lake of Embeddings index?

Accepted Answer

Currently, Lake of Embeddings can index multiple source types including Markdown directories (e.g., Obsidian vaults), specific SQLite database tables, and plaintext files. Future support for platforms like Notion and Confluence is planned.

Question 4

How does Lake of Embeddings integrate with Claude Code?

Accepted Answer

It integrates directly with Claude Code via an MCP (Meta-programming Capability Protocol) server. This allows you to seamlessly query your indexed knowledge base from within Claude Code, providing a powerful context for AI interactions and code generation.

Question 5

What is the primary benefit of using semantic search over traditional keyword search?

Accepted Answer

Semantic search finds content based on the underlying meaning and context, even if exact keywords aren't present. This allows for more relevant and comprehensive results, helping you discover connections and information in your personal knowledge base that a keyword search might miss.

Lake of Embeddings

Lake of Embeddings

主な機能

ユースケース

主な機能

ユースケース