MCP App Store
by KeenEthics
CaptainBook Statistics icon

CaptainBook Statistics MCP Server for ChatGPT

By CaptainBook

Overview

Analytics and reporting for tour and activity operators, answering questions about your own CaptainBook business data in plain language. Reports on revenue, bookings, products, resources, customers, sales channels, occupancy, extras, discounts and gift certificates. Ask which products sold best last season, how occupancy moved month over month, which sales channels brought revenue, or how this August compared with last. The connection is read-only. It cannot change a booking, a price or any setting, and it says so if you ask it to. Results are scoped to the business unit the connection was created for. Figures are historical. Asked a forward-looking question, it reasons from past patterns and tells you the answer is a heuristic rather than a forecast, because CaptainBook does not run forecasting models. Reporting distinguishes guests who travelled from bookings that were cancelled or expired, so attendance questions are answered from confirmed guests rather than raw totals.

How to connect

  1. ChatGPT: Open the CaptainBook Statistics listing in the ChatGPT app directory and choose Connect. ChatGPT listing CaptainBook Statistics exposes 5 tools here.

Tools

segment_fields_list

ChatGPT
List the FIELDS this Captainbook account's customer segments may filter on, and which comparison operators each field accepts. This is a VOCABULARY, not data: no customers, no segments, no counts — it is the dictionary you read BEFORE composing or editing a segment. For the audiences themselves call segments_list; for the people in one call segment_members. It is per-account and per-business-unit because the vocabulary is: custom attributes are operator-defined, and the behavioural fields exist only where customer-event tracking is switched on. So read it rather than assuming — a filter naming a field this account does not have is refused by the write, and a field you were never told about is an audience you cannot build. The operators matter as much as the names. not_contains on a behavioural field reads like a negation and is evaluated as "happened", so a tree written from the field name alone can select the exact opposite of what was asked for. An entry with operators: null constrains nothing and takes the shared set. Takes no arguments. Read-only; the write is inventory_segment_edit on the inventory server.

segment_get

ChatGPT
Read ONE customer segment in full by id: its name, description, type, lifecycle status, current member count, and the complete condition tree that defines who belongs to it. Use it to explain what an audience actually selects, and to get a WORKING conditions shape before composing a new segment — the filterable vocabulary is per-account, so copying a tree that already validates beats guessing field names (call segment_fields_list for the full vocabulary). It returns the segment DEFINITION, not the customers in it: for the people, call segment_members with the same id. Use segments_list to find the id — this tool takes one and does not search by name. A segment is a LIVE rule, not a saved list: membership is recomputed as customers qualify and stop qualifying, so the member count here is the count as of now. Read-only; editing a segment is inventory_segment_edit on the inventory server. A segment outside the business unit pinned to this MCP token reads as not found.

segment_members

ChatGPT
List the CUSTOMERS currently inside one segment, most recently entered first, paginated. This is the people tool: use it for "who is in my repeat-buyers audience", "give me that segment's contact list", or to sanity-check that a rule selects who the operator thinks it does. It takes exactly one segment id and will not list across segments — call segments_list for the ids, and segment_get when you want the RULE rather than the people. A segment is a LIVE rule, not a saved list: membership is recomputed as customers qualify and stop qualifying, so this is the answer as of now, it can differ tomorrow with no edit in between, and it is not a record of who was ever a member. A backfilling segment returns nothing yet because it has not finished computing — say so rather than reporting an empty audience. Rows are full customer records including names and contact details, so this needs permission to preview customers on top of permission to see segments: it can be refused for a user that segment_get answers fine for, and that refusal is about the PII, not about the segment. Read-only — it never adds or removes anyone. Scoped to the business unit pinned to this MCP token.

segments_list

ChatGPT
List the customer segments this Captainbook account has defined — the operator's saved audiences — paginated, each row carrying the segment's name, status, condition tree and CURRENT member count. This is the tool that answers "what audiences exist" and "which segment should I look at", and the one that gives you the id the other segment tools need. It returns the SEGMENTS, never the people in them: for the customers inside one, call segment_members with that segment's id. A segment is a LIVE rule, not a saved list: membership is recomputed as customers qualify and stop qualifying, so every members_count here is the count as of now and may differ tomorrow with no edit in between. A segment created moments ago reads backfilling with a count of 0 until its first recompute finishes — report that as "still filling in", not as an empty audience. Read-only; composing or editing a segment is inventory_segment_edit on the inventory server. Scoped to the business unit pinned to this MCP token.

statistics_query

ChatGPT
Query Captainbook statistics. Pick a metric (revenue, bookings, products, resources, locations, occupancy, etc.) plus date range and optional filters. Returns historical aggregates the assistant can narrate or compose with other calls. Read-only. Results are scoped to the business unit pinned to this MCP token.

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/captainbook-statistics"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/captainbook-statistics)