MCP App Store
by KeenEthics
Gainsight (Customer Communities) icon

Gainsight (Customer Communities) MCP Server for Claude and ChatGPT

By Gainsight

Overview

The Gainsight Community MCP connects your Customer Community to Claude. Surface unanswered questions, identify content gaps, and pull engagement trends through natural language. Review moderation queues, browse leaderboards, and audit community structure without waiting on a custom export. Purpose-built for community managers who'd rather ask their AI than navigate multiple dashboards.

How to connect

  1. Claude: Open the Gainsight (Customer Communities) listing in the Claude connectors directory and choose Connect. Claude listing Gainsight (Customer Communities) exposes 13 tools here.
  2. ChatGPT: Open the Gainsight (Customer Communities) listing in the ChatGPT app directory and choose Connect. ChatGPT listing Gainsight (Customer Communities) exposes 41 tools here.

Tools

admin_add_idea_tags

ChatGPT
Add tags to an idea.

admin_ask_question

ChatGPT
Ask a new question as yourself. Every call creates a new question.

admin_assign_idea_status

ChatGPT
Set an idea's status, replacing the current one. Re-assigning the same status still records a status change.

admin_assign_moderator_to_idea

ChatGPT
Assign a moderator to one or more ideas.

admin_assign_moderator_to_topic

ChatGPT
Assign a moderator to one or more articles, conversations, questions or product updates of one type. Send only the ID list that matches content_type.

admin_community_structure

ChatGPT
Inspect community structure and content layout. Use 'get_category_tree' for a full nested category hierarchy, 'list_categories' to page through all categories, 'get_category' for a single category's details (requires category_id), 'list_content_in_category' to list topics within a category (requires category_id; filter by content_type), 'get_visible_topic_counts' for topic counts per visible section (supports modules and top_level_section_ids filters), or 'list_recent_topics' for the latest topics across the community. Requires mcp:admin scope.

admin_create_article

ChatGPT
Create a new article as yourself. It is created as a draft — publish it with admin_publish_article. Every call creates a new article.

admin_create_conversation

ChatGPT
Start a new conversation as yourself. Every call creates a new conversation.

admin_create_product_update

ChatGPT
Create a new product update as yourself. It is created as a draft — publish it with admin_publish_product_update. Every call creates a new product update.

admin_edit_idea_content

ChatGPT
Replace an idea's body.

admin_edit_idea_moderator_tags

ChatGPT
Replace an idea's moderator tags.

admin_edit_idea_product_areas

ChatGPT
Replace the product areas of an idea.

admin_edit_idea_title

ChatGPT
Replace an idea's title.

admin_edit_topic_content

ChatGPT
Replace the body of an article, conversation, question or product update.

admin_edit_topic_title

ChatGPT
Replace the title of an article, conversation, question or product update.

admin_list_idea_statuses

ChatGPT
List the community's idea statuses (id, name, type, default, visible, displayOrder). Use it to find the ideaStatusId for admin_assign_idea_status.

admin_list_product_areas

ChatGPT
List the community's product areas (id, name), optionally filtered by name prefix. Use it to find the product area IDs for admin_edit_idea_product_areas.

admin_merge_idea_votes

ChatGPT
Move all votes from one idea into another. Irreversible — the votes cannot be moved back.

admin_moderation_overview

ChatGPT
Inspect community moderation queues. Use 'list_trashed_content' to page through deleted/trashed posts, or 'list_labeled_content' to list posts with specific moderation labels (filter by label name via moderation_labels). Requires mcp:admin scope.

admin_publish_article

ChatGPT
Publish a draft article.

admin_publish_product_update

ChatGPT
Publish a draft product update.

admin_recognition_overview

ChatGPT
Community recognition and gamification data. Use 'list_leaderboard' for top users by points, 'get_user_leaderboard_position' for a specific user's rank (requires user_id), 'list_badges' to enumerate available badges, or 'get_user_points_history' for a user's points timeline (requires user_id; supports period, from_date, to_date). Requires mcp:admin scope.

admin_remove_idea_tags

ChatGPT
Remove tags from an idea.

admin_reply_to_idea

ChatGPT
Post a reply on an idea as yourself. Every call adds a new reply.

admin_reply_to_topic

ChatGPT
Post a reply on an article, conversation, question or product update as yourself. Every call adds a new reply.

admin_set_idea_closed

ChatGPT
Close or reopen an idea.

admin_set_topic_closed

ChatGPT
Close or reopen an article, conversation, question or product update.

admin_submit_idea

ChatGPT
Submit a new idea as yourself. Every call creates a new idea.

admin_unvote_idea

ChatGPT
Remove your vote from an idea. Requires mcp:admin scope.

admin_user_overview

ChatGPT
Look up community user details. Use 'get_user_by_id' to fetch a full user profile by user_id, 'get_user_by_field' to find a user by a specific profile field (e.g. field='email', value='user@example.com'), or 'list_user_roles' to enumerate all roles assigned to a user (requires user_id). Requires mcp:admin scope.

admin_vote_idea

ChatGPT
Vote for an idea as yourself.

analytics_audience

ChatGPT
The ONLY tool for audience analytics, website traffic, pageviews, visitors, visitor segmentation, and visitor profiles. Admin-only tool. Requires authenticated access with administrator or community manager role. Use for ANY question about 'how many people visited', 'pageviews', 'traffic trends', 'unique visitors', 'registered vs guest visitors', 'audience size', 'daily traffic', 'visitor breakdown by section', 'who is visiting', 'audience segments', 'contributors vs viewers vs reactors', 'homepage traffic', 'visitor profiles', 'who visited and what did they do', or 'visitor engagement'. Answers: 'how many pageviews last month', 'what is the daily traffic trend', 'how many unique visitors this week', 'what percentage of visitors are registered', 'break down traffic by section', 'break down pageviews by content type', 'how many contributors vs viewers', 'show me audience segments', 'traffic to articles vs questions', 'homepage pageviews', 'show me visitor profiles for last quarter', 'which visitors created the most topics', 'who visited and earned the most points'. Use the 'view' parameter: 'summary' (default) for aggregate KPIs (total pageviews, unique visitors, registered/guest split, daily averages) with optional period-over-period comparison, 'daily_trends' for day-by-day pageview and visitor counts over time, 'pageviews_breakdown' for pageviews broken down by a chosen dimension (section, entity_type, content_type, or category_id), 'user_segments' for classifying active registered users into contributor/reactor/viewer tiers based on their activity, 'visitor_profiles' for registered visitor profiles with period-specific engagement stats (topics, replies, likes, logins) and gamification points — one row per visitor, sorted by activity. Do NOT use analytics_community_health for detailed audience or traffic questions — it only provides a high-level traffic summary. Do NOT use analytics_user_trends for visitor or pageview data — it handles user rankings and registration, not traffic. Do NOT use analytics_individual_user_profile for bulk visitor profiles — use visitor_profiles view instead. Do NOT use search tools for traffic or audience questions. Response notes: 'unique_visitors' = unique visitors by tracking cookie — different devices count as different visitors. 'registered_visitors' = unique registered users by user_id — different devices count as one user. 'guest_visitors' = visitors without a user account, cookie-based. 'pageviews' = total pages loaded or reloaded, not unique. 'registered_pageviews' / 'guest_pageviews' = pageviews split by visitor type. Guest pageviews are retroactively attributed to registered users when they later log in (matched by cookie). User segments: 'contributors' = users who created a topic or reply. 'reactors' = users who liked or voted but did not post. 'viewers' = users who only logged in. These are mutually exclusive tiers — a contributor is not also counted as a reactor.

analytics_community_health

ChatGPT
Cross-domain community health snapshot with period-over-period comparison. Admin-only tool. Requires authenticated access with administrator or community manager role. Returns a HIGH-LEVEL summary across 6 domains: users (total, new, active), content (topics created, replies), traffic (pageviews, unique visitors), Q&A (answer rate), support deflection, and gamification (badges, points). Use ONLY for broad 'how is the community doing' questions or when the user wants a single combined overview across multiple domains. For DETAILED analysis of any single domain, use the specialized tool instead: analytics_content_performance for detailed Q&A metrics/content gaps/response times, analytics_user_trends for user rankings/registration/churn, analytics_audience for detailed traffic/visitor analysis, analytics_individual_user_profile for a specific person. This tool gives breadth, the others give depth. Response notes: 'topic' = original thread (question, article, idea, conversation/discussion, product_update). 'reply' = response to a topic or another reply (threaded). 'post' = topic + reply combined. 'content_type' values: question, article, idea, conversation (also called discussion), product_update. All metrics include both current and previous period values. '_current' suffix = current period value. '_previous' suffix = previous period value. Previous period = same number of days immediately before the selected period. 'total_registered' = lifetime registered users excluding banned/deleted/guest/super-admin/not-activated/requires-approval. 'active_users' = users with any activity (post, reply, like, login) in the period. 'active_contributors' = users who posted, replied, liked, or voted (excludes login-only). 'avg_dau' = average daily active users. 'new_topics' = original threads created, not replies. For articles and product_updates, date filtering uses published_at (when content went live). Unpublished drafts are excluded. 'new_replies' = responses created. 'new_questions' / 'new_questions_answered' = questions created and answered in the period. 'answered_pct' = all-time % of questions with a best answer. 'page_views' = total pages loaded or reloaded. 'badges_earned' = badges awarded in the period. 'rank_promotions' = users promoted to a higher rank. 'points_awarded' = gamification points given in the period. 'primary_role_name' is a system ID — present 'primary_role_label' for display.

analytics_content_performance

ChatGPT
The ONLY tool for analysing topics, replies, and content performance across the community. Admin-only tool. Requires authenticated access with administrator or community manager role. Also tracks likes trends — how many likes were added in a period and which topics received the most likes. Use for ANY question about content metrics, top performing topics, content distribution, content gaps, response times, helpfulness scores, Q&A performance, content subscriptions, content creation patterns, or likes trends. Answers: 'what are the most viewed topics', 'how is content distributed across categories', 'which questions are unanswered', 'what is the average response time', 'show me content gaps', 'top topics by replies', 'helpfulness scores by category', 'Q&A answer rate', 'how many questions were answered', 'average time to answer', 'who is asking and answering questions by role', 'which content gets the most subscriptions', 'content creation by role', 'what did moderators post', 'how many likes were added last month', 'most liked topics'. Can filter by content type, category, group, group type, product area, tags, moderator tags, and moderation labels. For Q&A questions (answer rate, unanswered questions, response time) — ALWAYS use this tool, not analytics_community_health. For 'which content gets subscriptions' — use this tool. For 'which users subscribed to what' — use analytics_user_trends instead. For content performance data (views, replies, likes, helpfulness, response times) — use this tool, not search_unified (it searches content text but cannot return metrics). Do NOT use analytics_product_ideas for general content questions — that tool is ONLY for ideas and feature requests with vote counts. Response notes: 'topic' = original thread (question, article, idea, conversation/discussion, product_update). 'reply' = response to a topic. 'post' = topic + reply combined. 'content_type' values: question, article, idea, conversation, product_update. Date filtering for articles and product_updates uses published_at (when the content went live) instead of created_at; unpublished drafts are excluded. Questions, ideas, and conversations continue to use created_at. Reply date filters always use the reply's created_at. 'total_replies_created' = actual reply count, not topics-with-replies. 'unanswered_rate' = % of questions without a best answer. 'avg_first_reply_time_minutes' = average minutes from topic creation to first reply. 'helpful_votes_percentage' = % positive helpfulness votes (0-100). NULL = no votes. 0 = all votes were not-helpful. 'like_count' on a topic = likes on the topic itself, not its replies. 'subscription_count' on a topic = all-time subscriber count, not time-windowed. 'net_subscriptions' = subscribed minus unsubscribed in the period. 'answer_rate_pct' = questions_answered / questions_asked × 100. 'avg_answer_time_minutes' = average minutes from question creation to when the answer was marked as best. 'questions_answered' = questions with a best answer, not just any reply. 'unanswered' = questions with no best answer marked. 'gap_type' severity: unanswered = no best answer, zero_reply = no replies at all, high_view_low_reply = many views but few replies. 'primary_role_name' is a system ID — present 'primary_role_label' for display. 'moderator_tag_names', 'moderation_label_name', 'public_tag_names', 'product_area_names' are comma-separated when multiple values exist. qa_by_role returns two groups: 'askers' (who asks questions by role) and 'answerers' (who provides best answers by role). by_role 'role_type' = 'primary' for system roles, 'custom' for custom roles. response_time view returns avg_minutes = average first-reply time in minutes. sample_size = number of topics used in the calculation.

analytics_individual_user_profile

ChatGPT
Deep-dive into ONE specific named community user. Admin-only tool. Requires authenticated access with administrator or community manager role. ONLY use when the question names or identifies a specific person (e.g. 'tell me about John Smith', 'show me Sarah's profile', 'what has jsmith posted'). Look up by name or email to view their full profile (identity, roles, rank, company, groups, badges, lifetime stats, subscriptions), recent activity feed, gamification history, content performance, or group membership timeline. Use the 'view' parameter to select: 'profile' (default) for full profile, 'activity' for recent actions in a time window, 'gamification' for badge/rank/points history, 'content' for topics, replies, and content performance metrics, 'groups' for group membership timeline including joins, leaves, and role changes. IMPORTANT: For questions about whether a user was active, how engaged they were, or any complete activity summary, call this tool TWICE — once with view='activity' (for posts, likes, logins) and once with view='gamification' (for points earned, badges, rank changes). Activity view does NOT include points data. Do NOT use this tool for ranking, comparing, or listing multiple users — use analytics_user_trends instead. Do NOT use this tool for 'who has the most points', 'top contributors', 'most active users' — those are ranking questions that belong to analytics_user_trends. Response notes: 'topic' = original thread (question, article, idea, conversation/discussion, product_update). 'reply' = response to a topic. 'post' = topic + reply combined. 'content_type' values: question, article, idea, conversation, product_update. Date filtering in the content view for articles and product_updates uses published_at (when the content went live) instead of created_at; unpublished drafts are excluded. Questions, ideas, and conversations continue to use created_at. Reply date filters always use the reply's created_at. 'primary_role_name' is a system ID like 'roles.member' — present 'primary_role_label' instead (e.g. 'Member'). 'custom_role_names' is comma-separated. 'solved_count' = replies accepted as best answer, not closed topics. 'likes_given_count' = likes this user gave to others. 'likes_received_count' = likes this user received from others. 'vote_count' = net idea votes (upvotes minus unvotes), only for ideas. 'helpful_votes_percentage' = % positive helpfulness votes (0-100). NULL = no votes cast. 0 = all votes were not-helpful. 'like_count' on a topic = likes on the topic itself, not its replies. 'idea_status_type' values: open, closed, delivered, null (no status set). Activity view 'activity_type' values: topic_created, reply_created, topic_liked, reply_liked, topic_voted, enduser_login, sso_login, role_change. 'trigger_activity' in points breakdown shows the earning action (e.g. 'Created.topic', 'Created.reply').

analytics_moderation_activities

ChatGPT
Admin-only tool. Requires authenticated access with administrator or community manager role. Analytics on what moderators and admins do, in ONE tool with two targets selected by the 'target' parameter. target='content' (default): moderator actions on community CONTENT (topics, replies, private messages, DSA notices) - removals, spam, reports, labels, assignments, category moves, best answers. target='user': moderator/admin actions on USER ACCOUNTS - bans, GDPR erasures, role/seat changes, group membership decisions, segments, badges/ranks. Use it for questions like moderator/admin workload and activity, human vs AI vs system actions, volume over time, report resolution, ban/unban history, and the full history of one topic, reply or user. One row = one action by a moderator/admin or by the system (executed_by NULL or '0'); ordinary members' own actions are mostly not present. actor_type is human, system or ai (the ai-moderator account). Views summary, actor_split, activity_breakdown, daily_volume and time_distribution work for both targets; other views belong to one target. When 'target' is omitted it is inferred from 'view' (shared views default to 'content'). Filters for one target are ignored or rejected for the other. The 'view' parameter lists every view; 'summary' is the default. Time-windowed views return date_range and previous_period (default last 30 days). Notes: reports filed by ordinary members are not counted (report_funnel reflects reports filed by moderators or the system). Bans are DETECTED from subject_is_banned flipping to true across identity_access_user_updated events (there is no dedicated ban event). Many identity_access_user_updated / role_changed events are users editing their own account - compare executed_by with subject_user_id before calling something an admin action on another user. Not covered here: content performance (views, likes, answers) - use analytics_content_performance; user rankings/registrations - use analytics_user_trends; a profile of one named person - use analytics_individual_user_profile.

analytics_product_ideas

ChatGPT
The ONLY tool that can list, rank, sort, or count ideas by votes, views, or replies. Admin-only tool. Requires authenticated access with administrator or community manager role. Also tracks voting trends — how many votes were added in a period and which ideas gained the most votes. Use for any question about ideas, feature requests, product feedback, or the idea pipeline. Answers: 'which ideas have the most votes', 'top voted ideas', 'how many ideas were delivered', 'what feature requests are open', 'show me delivered ideas', 'idea pipeline stats', 'ideas by product area', 'how many votes were added last month', 'voting trends', 'which ideas gained votes', 'what percentage of votes are on delivered ideas', 'votes on delivered ideas this quarter'. Can filter by idea status (open/closed/delivered), product area, category, and minimum votes. Can sort by votes, replies, views, likes, or recency. Supports period-over-period comparison. For idea voting data, rankings, or vote counts — use this tool, not search_unified or search_discussions (they search idea text but cannot sort or rank by votes). Response notes: 'idea' = a feature request or suggestion submitted by a community user, always content_type='idea'. 'vote_count' = net idea votes (upvotes minus unvotes/merges/resets) — can be zero or negative. 'idea_status_type' values: open, closed, delivered, or null (no status assigned). 'ideas_no_status' = ideas with no status set. ideas_delivered + ideas_open + ideas_closed + ideas_no_status = ideas_count. 'idea_status_changed_at' = when status was last changed, null if never changed. 'date_filter_by' controls which date is filtered: 'created' = idea creation date, 'status_changed' = when status last changed. Delivered view defaults to 'status_changed'. 'avg_votes_per_idea' includes ideas with zero votes. 'product_area_names' is comma-separated when an idea belongs to multiple product areas. 'primary_role_name' is a system ID — present 'primary_role_label' for display. 'votes_delivered' = total votes on delivered ideas. 'votes_open' = total votes on open ideas. 'pct_votes_delivered' = votes on delivered ideas as % of delivered + open votes (excludes closed).

search_analytics

ChatGPT
Get search analytics data for the community. Provides insights into how users search: top queries, no-result searches, click-through rates, and more. Calls the syncsearch analytics API. Admin-only tool. Requires authenticated access with administrator or community manager role. Args: metric: Analytics metric. Valid values: top_searches, search_count, no_result_rate, click_through_rate, conversion_rate, frequent_searches_without_result, top_search_results, searches_without_clicks, no_click_rate, user_count, top_countries. Default: top_searches. date_from: Start date (YYYY-MM-DD, e.g. "2025-01-01"). date_to: End date (YYYY-MM-DD, e.g. "2025-12-31"). limit: Maximum results (default 10, max 100).

search_discussions

ChatGPT
Find active community discussions on a topic. Returns questions and discussions with status, reply counts, and engagement metrics.

search_unified

ChatGPT
Search across all content sources via the Algolia unified index. Searches community posts, knowledge base articles, Zendesk tickets, Salesforce articles, Freshdesk articles, and Skilljar courses in one query.

admin_community_structure

Claude

admin_moderation_overview

Claude

admin_recognition_overview

Claude

admin_user_overview

Claude

analytics_audience

Claude

analytics_community_health

Claude

analytics_content_performance

Claude

analytics_individual_user_profile

Claude

analytics_product_ideas

Claude

search_analytics

Claude

search_discussions

Claude

search_unified

Claude

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/gainsight-customer-communities"><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/gainsight-customer-communities)