Yes. Rixe exposes a REST API, a Looker Studio connector and an MCP server. The MCP server lets an AI assistant read a brand's visibility data directly, so the metrics can be queried in the same place the rest of the reporting already happens.
The API is the plain path: authenticate, request a project's scores, mentions, ranks and citations, and pull them into whatever warehouse or dashboard already holds the rest of your reporting. The Looker Studio connector exists so the same data can land in a client report without anyone writing code.
The MCP server is the more interesting one. MCP lets an AI assistant read a tool's data directly, so a question like "which prompts did we lose this week and which source is behind it" can be answered against live Rixe data, in the same conversation where the rest of the work is happening. It removes the step where someone exports a CSV to answer a question that was already a sentence.
All three read the same underlying records, so a number in a client deck, a number in a dashboard and a number an assistant quotes cannot disagree with each other.
Three ways in, for three jobs
The REST API is the plain path: authenticate, request a workspace's scores, mentions, ranks and citations, and pull them into whatever warehouse or dashboard already holds the rest of your reporting. It is read-only, versioned and paginated, which is the boring set of properties a data team actually asks about.
The Looker Studio connector exists for the case where nobody wants to write code: point it at a workspace, pick the fields, and the same numbers land in a report someone else in the company already knows how to edit. It reads the same endpoints the API exposes, so the two can never disagree.
What the MCP server changes
The MCP server lets an assistant read the data directly, which turns a dashboard question into a conversation. Instead of opening four screens to find out why last week moved, you ask — and the assistant reads the same measurements, with the same sample sizes attached, and answers from them rather than from a guess.
It is read-only by design. An assistant can pull a score, a list of mentions or the sources behind an answer; it cannot change a prompt set, approve a draft or publish anything. The boundary is deliberate: reading is where an assistant is useful, and writing is where an audit trail matters more than convenience.
Connections that are not an API
Most teams never touch the API and still connect three things: Search Console and Analytics on the way in, so clicks and sessions sit beside the visibility numbers, and a CMS on the way out, so an approved draft becomes a draft post rather than a copy-paste. All three are requested with the narrowest scope that does the job.
Nothing writes to your site without a person pressing approve, and the CMS connection creates a draft rather than a published post — a deliberate limitation that has annoyed exactly nobody in practice, because the review step is where a claim without a source gets caught.
