MCP tool reference

The Profinity Model Context Protocol (MCP) server provides ten read-only tools for components, tags, tag history, and alerts. Your AI tool chooses and calls them when you ask it a question, so you do not call them yourself. The table shows what each tool answers and the permission it needs, so you can tell which question a missing permission blocks. Every tool also needs MCP integration.

Tool	What it answers	Permission
get_all_components	Which components are in the active profile	View tags
search_tags	Which tags match a name, unit, or component, up to 50 at a time	View tags
get_tag_catalog	Which tags a component or tag path holds	View tags
get_tag_sample_by_canonical_id	The current value of one tag	View tags
get_tag_samples	The current values of up to 50 tags	View tags
get_tag_series	The values of one tag over a time range, summarised by max (the default), min, mean, sum, count, last, or first	View tags
get_tag_series_batch	The values of up to 20 tags over the same time range	View tags
get_alert_rules	Which alert rules are defined, and whether each is enabled	View alerts
get_active_alerts	Which alerts are active now, and whether each is acknowledged or silenced	View alerts
get_alert_history	When alerts were raised, cleared, acknowledged, or silenced	View alerts

None of the tools acknowledge, silence, or otherwise change an alert or a tag. Make those changes in the Profinity UI or through the REST API.

Ask about one component, or a narrow time range, when you can. A tool refuses a whole-profile tag list that is too large to return in one answer, and the AI tool then has to ask again with a narrower scope.

The answer silently includes only the tags and alerts that the account can view. When a question returns less than you expect, refer to Fewer results than expected. To connect a tool, refer to Prepare Profinity for an MCP client.