Back to Apps

kvault
by rabrooks
Local-first searchable knowledge corpus with BM25 ranking and MCP server integration.
0 stars
Works in:CursorWindsurf
Exposes:ToolsResources
What it does
kvault is a fast, local-first knowledge base that stores information as Markdown, JSON, or plain text. It allows users to build a personal or team knowledge corpus with powerful search capabilities, including ripgrep for speed and Tantivy for BM25 ranked fuzzy matching, without relying on external AI embeddings APIs.
Tools
search_knowledge: Search the corpus for matching documents using keyword or BM25 ranking.list_knowledge: List all documents, with optional category filtering.get_document: Retrieve the full content of a specific document by its path.add_knowledge: Programmatically add new documents to the knowledge base.
Installation
Install ripgrep first, then install kvault:
cargo install kvault
To enable MCP support, build with the mcp feature:
cargo build --release --features mcp
Claude Desktop config:
{
"mcpServers": {
"kvault": {
"command": "kvault",
"args": ["serve"]
}
}
}
Supported hosts
- claude
- cursor
- windsurf
Quick install
cargo install kvaultInformation
- Pricing
- free
- Published
- 6/29/2026
- stars
- 0
Categories
Choose your AI client and follow the steps below.
Cursor
Add to mcp.json: {"kvault": {"command": "kvault", "args": ["serve"]}}Windsurf
Add to mcp_config.json: {"kvault": {"command": "kvault", "args": ["serve"]}}Claude Desktop
{"mcpServers": {"kvault": {"command": "kvault", "args": ["serve"]}}}





