mcp-server and netbox-mcp-server
These tools are competitors, as both implement the Model Context Protocol (MCP) server, but for distinct platforms: the first for Keboola and the second for NetBox, making them mutually exclusive choices depending on the target data source for LLMs.
About mcp-server
keboola/mcp-server
Model Context Protocol (MCP) Server for the Keboola Platform
Exposes Keboola's data pipelines, transformations, and job orchestration as composable tools via the Model Context Protocol, enabling AI agents to query tables, create SQL transformations, trigger jobs, and manage development branches through a unified interface. Supports both a hosted remote server with OAuth authentication (for Cursor, Claude, Windsurf, and other MCP clients) and local deployment with configurable tool authorization via HTTP headers for access control. Built on Streamable HTTP transport with granular read-only and tool-filtering capabilities to safely integrate Keboola workflows into AI agent workflows.
About netbox-mcp-server
netboxlabs/netbox-mcp-server
Model Context Protocol (MCP) server for read-only interaction with NetBox data in LLMs
Exposes three core tools—`get_objects`, `get_object_by_id`, and `get_changelogs`—for querying NetBox infrastructure data and audit trails directly within LLM conversations. Supports both stdio and HTTP transports, integrating seamlessly with Claude Desktop/Code and other MCP-compatible clients via environment variable configuration. Includes optional field filtering to reduce token usage by up to 90% when querying large datasets.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work