MCP App Store
by KeenEthics
GatherHQ icon

GatherHQ MCP Server for Claude and ChatGPT

By GatherHQ

Overview

GatherHQ connects Claude to your voice-of-customer research through Emma, Gather's AI research teammate. Ask in plain language and Emma scopes and launches real research studies (proposing a research plan and interview guide for your approval before fielding), pulls findings, quotes, and insights from completed studies, chats with buyer personas grounded in real interview responses, and turns results into content like reports, decks, blog posts, and ad creatives. Emma keeps context across the conversation, so follow-ups and approvals work naturally. Built for marketing and research teams who want customer evidence on demand instead of waiting weeks for a study.

How to connect

  1. Claude: Open the GatherHQ listing in the Claude connectors directory and choose Connect. Claude listing GatherHQ exposes 1 tool here.
  2. ChatGPT: Open the GatherHQ listing in the ChatGPT app directory and choose Connect. ChatGPT listing GatherHQ exposes 22 tools here.

Tools

gather_ask_persona

ChatGPT
Ask one of your workspace's buyer personas a question and get its answer in its own voice, grounded in the research behind it. Answers are simulated, not quotes from real people. When its research does not cover the question, the persona says so and Gather logs the question to the persona's research backlog. Get persona_key from gather_list_personas.

gather_create_content

ChatGPT
Create a report, blog post, ebook, slide deck or social card from a completed Gather study with real respondents, and return its link. Put the angle, audience and framing in brief.

gather_create_embed_session

ChatGPT
Create one short-lived, single-use sign-in ticket for the connected person and workspace. Called only by the GatherHQ interface.

gather_creative_test_check

ChatGPT
Use this to follow a creative test started with gather_test_creatives, with its id. While it runs, it returns the status; when it is done: a predicted grade per creative against the others (A front-runner, B in the running, C unlikely to win), what to keep, rework or retire and why, what to test next, and simulated reactions. Call it about once a minute until it is complete or failed. It starts nothing.

gather_design_study

ChatGPT
Design a study with real people, one step at a time. The first call (no study_id) starts a new design and creates a draft study in your workspace; each step's answer returns study_id, step and what to do next. Pass that study_id on every later call to continue the same design. step "questions": Gather asks for something missing (objective, audience, who qualifies); ask the person as written and send their answer in instructions. step "plan_ready": Gather shares a link to a research plan and interview guide (Gather's research team sees the plan too); send feedback in instructions to revise it, or, only when the person approves it, call again with approve true. Gather then builds the interview, and its audience targeting goes to Gather's research team: the answer says so and has no test link yet, so check on the study later with gather_list_studies. (For Gather staff, Gather sets up the audience at once and returns a test link.) step "built": done; nothing goes to real people until gather_launch_study.

gather_export_responses

ChatGPT
Export a study's conversations as a CSV and return a private download link that works for 60 minutes. Completed respondents only unless include_incomplete is true.

gather_find_quotes

ChatGPT
Search or browse your verbatim library: word-for-word quotes from real respondents in your completed Gather studies, with the question asked, themes, sentiment and study. Leave search out to browse. Quote people exactly and never reword a quote into a new claim.

gather_find_stats

ChatGPT
Use this when someone wants a stat for an outbound email, an ad or a slide, e.g. "a stat for my email to a grocery category manager": returns the workspace's stats that fit, best first. Each is a stat and headline from one of its finished social cards (studies with real people), with its source line (e.g. "Real interviews · 48 people · Nov 2026"), the study and a link to the card. Pass the prospect's industry, role and topic; leave them out for the newest stats. Give each stat and headline word for word with its source line beside it; never change a number or add a claim. For quotes use gather_find_quotes; to answer a question use gather_research_answer.

gather_get_simulation

ChatGPT
Get the status or results of a simulation started with gather_run_simulation: the summary, each persona's direction and reason, and a link. While it is still running, this waits up to about 3 minutes for the results.

gather_launch_study

ChatGPT
Launch a built study to real people: Gather starts recruiting its audience and the interviews go live. This cannot be undone and uses the workspace's Gather credits. Call it only after the person explicitly says to launch this study. Needs a workspace with Gather credits. A study whose only audience is a manual link recruits nobody: when its design here named who to recruit, Gather tries to set up that targeting first and does not launch if it can't. Some studies go to Gather's team for review before they go live.

gather_list_content

ChatGPT
Link your workspace's core research assets (messaging house, positioning, buyer personas, competitive framing, brand health, verbatim library) and the content made from your studies (reports, blogs, decks), and say which core assets do not exist yet. Core assets build when a study with real respondents completes.

gather_list_personas

ChatGPT
List the buyer personas in your Gather workspace, each with the persona_key gather_ask_persona needs. Personas are built from your research and answer as simulations.

gather_list_studies

ChatGPT
List your Gather workspace's 20 newest studies, each with its study_id and status (draft, live (in field), in review or completed); search and status narrow those 20. Use for "what's running?", a status update, or to find the study_id another tool needs.

gather_report_check

ChatGPT
Get the status of a report started with gather_report_start. When completed, returns the headline, findings, option choice, themes, quotes and the report link.

gather_report_list

ChatGPT
List the reports started with this connection, newest first, with their report links.

gather_report_start

ChatGPT
Start a customer research report for a company. Gather researches the company, designs the interview guide, runs simulated interviews with the target audience (default 30) on its main research engine, and writes a branded report. Returns an id and a report link immediately (the link shows live progress); then call gather_report_check about once a minute. Reports with 30 interviews take about 16 minutes.

gather_research_answer

ChatGPT
Answer a question about your customers from your own Gather research: your completed studies and the files you added to Gather. Returns counted numbers, word-for-word quotes, what to do and a confidence; when your research does not cover the question, it says so. Use for questions like "What would make CIOs switch?" or "Why do customers choose us?". Ask one standalone question that names the group instead of "they". Your question and the start of the answer are saved to your workspace. Share the answer as written. For quotes alone use gather_find_quotes; for a study's status use gather_list_studies.

gather_research_ideas

ChatGPT
Suggest research ideas for your company, each with why now and supporting signals with their sources. Use when someone asks what to research next. When no recent ideas are saved, Gather looks up current signals about your company on the web and saves new ones to your workspace.

gather_run_simulation

ChatGPT
Start a simulation of how your workspace's saved personas would react to a change or idea (a price change, a message, a feature), on Gather's simulation engine, and wait up to about 3 minutes for the results: the summary, each persona's direction and reason, and a link. If it is still running, get the results later with gather_get_simulation. Results are simulated, not observed customer responses. Needs a workspace with Gather credits; otherwise use gather_report_start, which runs simulated interviews for any company.

gather_set_study_audience

ChatGPT
Send who a draft study should recruit, from a description such as "CTOs at B2B SaaS companies with 50 to 500 employees". Gather's research team sets up the targeting and its screener questions from it: the answer says the team has it, with no test link yet. (For Gather staff, Gather sets it up at once and returns a test link.) A live study's audience cannot change from here.

gather_test_creatives

ChatGPT
Use this when someone wants their target buyers' reactions to ad creatives or other images before they run them: which of up to 5 lands best, and why. Each image is shown as a sponsored post in a feed to simulated respondents from the workspace's personas (or the audience described). Send each image as an https url (recommended) or small base64 data: JPEG, PNG, WebP or GIF, up to 5 MB each; video is not supported. Returns an id at once; call gather_creative_test_check about once a minute (a few minutes in all). Results are simulated and directional: a predicted ranking against each other with reasons, not click-through rates or live ad performance. Needs a Gather workspace with a contract or Gather credits, and counts as one run. Do not use it to write ad copy or make images; for text options such as headlines or taglines, use gather_report_start.

ask_emma

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/gatherhq"><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/gatherhq)