ck
Local first semantic and hybrid BM25 grep / search tool for use by AI and humans!
ck (seek) is a local-first semantic code search tool that functions as both a command-line utility and an MCP server for AI agent integration. It enables AI assistants and developers to find code by meaning rather than exact keywords, using local embedding models to understand concepts like "error handling" or "authentication logic" without those exact terms being present. The tool runs completely offline with FastEmbed-powered semantic search, traditional regex/grep capabilities, and hybrid search combining both approaches. ck integrates with Claude Desktop, Cursor, and any MCP-compatible client through its built-in MCP server mode.
Key Features
Use Cases
- 01AI-assisted code exploration where agents search codebases by concept ("find retry logic", "locate authentication flows")
- 02Security audits combining semantic and keyword search for sensitive patterns (credentials, input validation)
- 03Code review preparation by identifying refactoring candidates, duplicate logic, or untested source files
- 04Documentation generation by finding public API surfaces and related code sections
- 05Team onboarding by semantically locating feature implementations without knowing exact variable names
- 06CI/CD integration for automated scanning of code changes for TODOs, vulnerabilities, or missing tests
Related MCP Servers
View moremarkitdown
Python tool for converting files and office documents to Markdown.
firecrawl
The context API to search, scrape, and interact with the web at scale. 🔥
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
ck — FAQ
What is ck and how does it work with AI agents?+
ck is a semantic code search tool that provides an MCP server (via --serve flag) exposing search capabilities to Claude Desktop, Cursor, and other MCP-compatible AI clients. It runs local embedding models to understand code by meaning, not just keywords, and operates completely offline without external API calls.
How do I install ck for use with Claude Desktop?+
Install via cargo install ck-search, then add it to Claude Desktop's MCP server configuration by running 'claude mcp add ck-search -s user -- ck --serve' or manually editing the config JSON. You'll need to approve tool permissions when prompted.
Which AI clients and editors support ck?+
ck works with Claude Desktop, Cursor, and any MCP-compatible AI client through its MCP server mode. It can also be used standalone via CLI or its interactive TUI from any terminal.
Does ck require API keys or internet access?+
No. ck runs completely offline using locally cached FastEmbed models with no network calls or API keys required. Models are downloaded once to ~/.cache/ck/models/ and reused.
Is ck free and open source?+
Yes, ck is open source under MIT/Apache-2.0 dual license and free to use. It's available on crates.io and can be installed via cargo install ck-search.
What are the prerequisites for running ck?+
You need Rust 1.88+ to install from source or cargo. The tool automatically downloads and caches embedding models (~100-500MB) on first use. No other dependencies or external services are required.
How do I install ck?+
Open the source repository on GitHub and follow its README. ck is a mcp server — MCP Agents Market links you directly to the official repo.
Is ck free?+
ck is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.