codesearch MCP Server

Unified code search across 15+ sources with embedding dedup + cross-encoder reranking for OpenCode.

Tools

search_all — Unified search across 15+ sources (Context7, GitHub, DeepWiki, CodeWiki, SO, SOFA, HN, Libraries.io, npm, crates, DevDocs, Semantic Scholar, CORE API, arXiv, Tavily) with NIM dedup + BM25 + cross-encoder reranker
github — GitHub search (code/repos/issues/users) + repo ops (readme, contents, languages, topics, releases)
docs — Documentation smart search (Context7→ReadTheDocs→llms.txt→DevDocs) + direct DevDocs/ReadTheDocs access
search_package — Package metadata from npm/PyPI/crates + deps.dev dependency graphs
so_search — Stack Overflow via Stack Exchange API (free) or SOFA
hn — Hacker News search, item detail, story lists
search_libraries — Libraries.io dependency metadata
vulns — Sonatype Guide vulnerability scanning
papers — Semantic Scholar, CORE API, and arXiv academic paper search
wiki — Repo architecture via DeepWiki + CodeWiki

Name		Name	Last commit message	Last commit date
Latest commit History 37 Commits
.gitignore		.gitignore
README.md		README.md
__init__.py		__init__.py
code_expand.py		code_expand.py
codewiki.py		codewiki.py
config.py		config.py
context7.py		context7.py
core_api.py		core_api.py
deepwiki.py		deepwiki.py
depsdev.py		depsdev.py
devdocs.py		devdocs.py
embed.py		embed.py
enrich.py		enrich.py
githits.py		githits.py
github_api.py		github_api.py
hackernews.py		hackernews.py
libraries_io.py		libraries_io.py
oss_index.py		oss_index.py
pkg_utils.py		pkg_utils.py
readthedocs.py		readthedocs.py
registries.py		registries.py
requirements.txt		requirements.txt
reranker.py		reranker.py
resilience.py		resilience.py
run.example		run.example
semantic_scholar.py		semantic_scholar.py
server.py		server.py
sofa.py		sofa.py
stack_exchange.py		stack_exchange.py
tavily_search.py		tavily_search.py