Shebe FAQs

Question 1

What is Shebe?

Accepted Answer

Shebe is a high-performance BM25 full-text search engine specifically designed for code repositories. It integrates seamlessly with Claude Code via its 14 MCP tools, enabling developers to quickly find functions, APIs, and patterns across large codebases.

Question 2

How fast is Shebe for code search and indexing?

Accepted Answer

Shebe delivers exceptional speed with 2ms query latency for search results and impressive indexing rates of 2,000 to 12,000 files per second. This allows for near-instantaneous content discovery and rapid repository indexing.

Question 3

Does Shebe require a GPU or embeddings for its search functionality?

Accepted Answer

No, Shebe utilizes BM25-only search, meaning it operates efficiently without the need for embeddings or a GPU. This makes it a lightweight and accessible solution for fast code search.

Question 4

How does Shebe enhance AI-assisted development with Claude Code?

Accepted Answer

Shebe integrates with Claude Code through 14 specialized MCP (Meta-Coding Protocol) tools. These tools allow developers to perform context-aware content searches directly within Claude, significantly reducing token usage and speeding up tasks like refactoring or codebase exploration compared to traditional methods.

Question 5

What types of files and character sets does Shebe support?

Accepted Answer

Shebe offers full UTF-8 support, enabling it to handle diverse character sets, including emojis and CJK characters. It can search not only code files but also non-code files like YAML and Markdown, making it a versatile polyglot search solution for any repository.

Shebe

Shebe

주요 기능

사용 사례

주요 기능

사용 사례