HyperDB MCP Server (v0.4.0) is an MCP (Model Context Protocol) server that turns Hyper, a columnar SQL database, into an instant analytics engine for LLM workflows. It provides zero-setup SQL analytics, ingesting data from JSON, CSV, Parquet, or Arrow formats and enabling query via SQL. Built on pure-Rust HyperDB, it achieves high ingest and query throughput (22M+ rows/sec insert, 18M+ rows/sec query). The server supports persistent databases for cross-session memory, shared daemon for multiple MCP clients, read-only mode, and inline chart generation. As a pre-1.0 release (0.x), it may undergo breaking changes frequently. Release cadence is rapid with multiple versions per month. Key differentiators: optimized for AI toolchains, combines structured queryable memory with columnar analytics, and features incremental ingest via watch directories.
npm install hyperdb-mcpNo compatibility data collected yet for this library.
Verified import paths — ran on the pinned version, not inferred.
Initializes the HyperDB MCP server, connects via stdio transport, and performs a basic ingest-and-query operation on a CSV file.
Pin version to current minor if relying on stable API; watch releases for migration guides.
Use Node 21+ or use a transpiler like esbuild to convert ESM to CommonJS.
Set environment variable HYPER_PORT to a custom port or kill existing Hyper instance before launching.
Monitor changelog for updated incremental ingest API.
Install @modelcontextprotocol/sdk: npm install @modelcontextprotocol/sdk. For ESM, ensure import path ends with '.js'.
Add 'type':'module' to package.json and use import statements. Or use Node 21+ and ensure .mjs extension.
Set HYPER_PORT environment variable to an unused port. Check for existing Hyper processes: lsof -i :<port> and kill them.