searchcode vs Shebe

searchcode provides code-intelligence dossiers for public GitHub repositories, covering languages, complexity, technology stacks, code quality, and secrets. Its MCP server exposes six tools for AI assistants, and each tool is also available as JSON over HTTP through the REST API. Analysis reports language breakdowns, complexity metrics, detected technologies, quality findings, and credentials such as API keys, tokens, and private keys with file, line, and confidence details. Search supports full text, boolean logic, regular expressions, and structural filters for declarations, usages, strings, and comments. File retrieval can return complete files, line ranges, or named symbols with nearby declarations; batch retrieval handles up to 10 files. The free beta plan covers public GitHub repositories without signup, an API key, or installation. Enterprise deployment runs on customer infrastructure as a Docker image with a Kubernetes Helm chart. Private repositories can authenticate with SSH keys or HTTPS tokens.

Shebe is a local code-search tool that uses BM25 full-text search across code repositories. It ranks results and supports phrase searches, Boolean queries, and literal matching. Fourteen MCP tools serve coding agents with functions including search, repository indexing, file reading, and reference finding. Shebe runs without embeddings, a GPU, or a cloud service, but repositories must be cloned locally before indexing; remote repository URLs are not supported. Its text-based search does not provide semantic similarity, so related words such as “login” and “authenticate” may not match. Prebuilt binaries are listed for Linux x86_64 and macOS on Apple Silicon and Intel; Linux ARM64 is planned. Windows users need to build it from source. The installation guide covers MCP setup and Zed configuration and points to the organization for VS Code and other editor extensions. Shebe is free and licensed under Apache-2.0. The README reports 2 ms query latency and indexing rates from 1,928 to 11,210 files per second in its listed benchmarks.