Integrates with OpenAI's embedding models to enable semantic code search, allowing for hybrid ranking that combines traditional TF-IDF keyword matching with vector-based semantic understanding.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@CodeRAG find the implementation of the JWT token validation"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Locus gives coding agents fast, local code search that returns complete, implementation-ready chunks instead of grep dumps or whole-file guesses.
npx -y @sylphx/locus --root=/absolute/path/to/projectFor Claude Code:
claude mcp add locus -- npx -y @sylphx/locus --root=/absolute/path/to/projectAsk a natural-language question:
{
"query": "where is user authentication enforced?",
"limit": 5
}Locus returns ranked AST chunks with file paths, line ranges, symbol names, score explanations, freshness, and explicit gaps.
Related MCP server: mcplens
Ask your agent | Locus returns |
“Where is this behavior implemented?” |
|
“What code is related to this function?” |
|
“Search across these repositories.” | cross-repo search |
“Find the implementation, not tests or declarations.” | ranked code chunks |
“Give me enough context to edit safely.” | compact, deduplicated results |
Tool | Purpose |
| Hybrid local search over code chunks |
| Find related chunks from a known file and line |
The public surface stays small. Indexing is cached locally and refreshed when the repository changes. Optional embeddings are opt-in and never required for the default path.
Locus does not hide expensive work behind “auto”.
fast is the default local TF-IDF and AST path.
quality enables optional vectors and reranking when configured.
research is not a Locus mode; use a separate web tool for external sources.
No Docker, vector database, or API key is required.
Generated files, dependencies, and build output are excluded by default.
AST boundaries return functions, classes, and methods.
Hybrid ranking combines lexical and semantic evidence when available.
Results include score components instead of opaque ordering.
Token budgets and dedup keep context small.
Local indexes keep code on the machine.
Product | Job |
PDF answers with page-level proof | |
Image facts and pixel evidence | |
Video timelines and timestamp evidence | |
Repository architecture and impact | |
Web research with source excerpts |
Locus finds the code chunk; Spine maps the repository architecture. Each product is independent, so install only the tools your agent needs.
bun install
bun run build
bun test
cargo test
bun run benchmark:public-proof
bun run benchmark:release-gateMIT
This server cannot be deployed
An MCP server that gives your AI access to the source code and docs of all public github repos
Driflyte MCP server which lets AI assistants query topic-specific knowledge from web and GitHub.
Capability registry for the agentic economy. Semantic search over verified MCP server listings.
