Memento FAQs

Question 1

Which programming languages does Memento support for semantic analysis?

Accepted Answer

Memento currently provides language-aware semantic relationships and context retrieval for Go, TypeScript/JavaScript, and PHP. This helps AI agents understand the intricacies and relationships within code written in these languages.

Question 2

What is Memento and how does it benefit AI agents?

Accepted Answer

Memento is a local-first MCP server designed to provide AI agents with durable, high-signal memory for your code repositories. It offers fast semantic code context, allowing AI agents to better understand and interact with your codebase, improving their efficiency and accuracy.

Question 3

What specific repository operations can AI agents perform using Memento?

Accepted Answer

Through its exposed MCP tools, AI agents can perform operations such as listing repository files (`repo_list_files`), reading file content (`repo_read_file`), semantic searching (`repo_search`), finding related files (`repo_related_files`), and retrieving focused code context (`repo_context`). It also supports memory management like `memory_upsert` and `memory_search`.

Question 4

How does Memento store and retrieve memory for AI agents?

Accepted Answer

It maintains an on-disk code index per repository for fast, bounded context retrieval and stores explicit, durable, repo-scoped notes. These notes and indexed context persist across sessions, acting as reliable long-term memory for AI agents.

Question 5

Does Memento support incremental updates for code changes?

Accepted Answer

Yes, Memento supports incremental change detection. For Git repositories, it polls `git status` for a fast path, and for non-Git repositories, it uses a filesystem watcher as a fallback, ensuring its index is always up-to-date with your latest code.

Memento

Memento

主な機能

ユースケース

主な機能

ユースケース