Melchizedek FAQs

Question 1

How does Melchizedek enhance search for Claude Code?

Accepted Answer

It provides production-grade hybrid search, combining BM25 keyword matching with dual text/code vector embeddings, Reciprocal Rank Fusion (RRF), and a cross-encoder reranker for highly relevant results from your past conversations.

Question 2

What is Melchizedek?

Accepted Answer

Melchizedek is a developer tool that gives Claude Code persistent memory by automatically indexing all your conversations, enabling you to search across your entire coding history, projects, and debugging sessions.

Question 3

Can Melchizedek search across different projects?

Accepted Answer

Absolutely. Melchizedek indexes conversations from all your Claude Code projects, allowing you to seamlessly search and retrieve past solutions, architectural decisions, and code patterns regardless of which project you were working on at the time.

Question 4

How does Melchizedek optimize token usage during search?

Accepted Answer

It employs a progressive retrieval pattern: starting with compact snippets from 'm9k_search', offering more context with 'm9k_context', and providing full content with 'm9k_full'. This approach saves up to 4x tokens compared to loading everything at once.

Question 5

Is Melchizedek fully local and private?

Accepted Answer

Yes, Melchizedek operates 100% locally with zero configuration and no API keys required. Your data is stored securely in a single SQLite file on your machine, ensuring complete privacy.

Melchizedek

Melchizedek

主要功能

使用案例

主要功能

使用案例