Memex FAQs

Question 1

Which AI systems can integrate with Memex?

Accepted Answer

Memex is built specifically for Claude Code but is designed to be compatible with any client that adheres to the Model Context Protocol (MCP). It registers its tools (like `ask`, `memorize`) for seamless integration within these environments.

Question 2

What is Memex and its primary purpose for AI agents?

Accepted Answer

Memex provides AI agents with persistent, searchable semantic memory. It acts as a local Model Context Protocol (MCP) server, enabling agents to store, retrieve, and reflect on knowledge by meaning, ensuring they retain context across sessions and interactions.

Question 3

Can knowledge within Memex be organized or categorized?

Accepted Answer

Yes, Memex supports configurable knowledge categories such as `personal`, `preference`, `decision`, `project`, `routing`, and `reference`. This allows for structured memory management and filtered retrieval, making the agent's knowledge base more organized and efficient.

Question 4

How does Memex store and retrieve knowledge semantically?

Accepted Answer

Memex utilizes an embedded vector database (LanceDB) for knowledge storage. It automatically generates high-dimensional vector embeddings (using OpenAI's models) for all entries. This allows agents to perform semantic searches, finding relevant context based on meaning rather than exact keywords.

Question 5

What tools does Memex offer AI agents?

Accepted Answer

Memex exposes four core tools: `memorize` to store new knowledge, `ask` for semantic search and context retrieval, `forget` to remove specific entries, and `reflect` to audit or review what's stored about a particular topic.

Memex

Memex

主要功能

使用案例

主要功能

使用案例