FIWARE MCP Server

The Dncampo Fiware MCP Server is a lightweight bridge that connects AI assistants, such as Claude, to FIWARE’s Context Broker ecosystem. By exposing a set of MCP tools that wrap standard NGSI operations, the server lets conversational agents query, inspect, and manipulate contextual data without needing direct access to the broker’s HTTP API. This eliminates boilerplate code for developers and enables AI workflows that can react to real‑time sensor feeds, IoT telemetry, or smart city information sources.

At its core, the server implements three essential tools: CB_version, query_CB, and publish_to_CB. The CB_version tool simply reports the broker’s API version, which is useful for ensuring compatibility or for debugging. query_CB allows the assistant to perform arbitrary NGSI queries—filtering entities, selecting attributes, or retrieving historical data—by passing a query string and receiving the broker’s JSON payload. Finally, publish_to_CB handles both creation and update of entities, accepting a dictionary that follows the NGSI‑LD format and returning an operation status. These tools cover the most common CRUD patterns needed when integrating contextual data into AI‑driven decision making.

For developers, this MCP server offers several tangible advantages. It abstracts the low‑level NGSI protocol, reducing the cognitive load when designing AI prompts that need to read or write context. The server’s error handling is robust: it gracefully reports network issues, malformed data, or broker errors back to the assistant, allowing conversational flows to adapt or retry automatically. Moreover, the server is intentionally modular; its configuration (host, port, timeout) can be tweaked to match any FIWARE deployment—whether a local NGSI‑LD broker or a cloud‑hosted instance.

Typical use cases include smart building automation, where an AI assistant can query temperature or occupancy sensors and publish new control commands; urban mobility platforms that need to fetch real‑time traffic data before advising users; or industrial IoT dashboards that let operators ask for the status of a machine and have the assistant update maintenance schedules. In each scenario, the MCP server removes the need for custom API wrappers, enabling rapid prototyping and reliable production deployments.

By integrating seamlessly with existing MCP tooling frameworks (e.g., and commands), the Dncampo Fiware MCP Server fits naturally into AI development pipelines. It provides a single point of contact for context‑aware assistants, empowers developers to build richer conversational experiences, and opens the door to automated, data‑driven decision making across FIWARE ecosystems.