Hubble MCP connects AI assistants such as ChatGPT and Claude directly to your Hubble workspace. Instead of manually copying study plans, transcripts, responses, or findings into an AI tool, you can ask your AI assistant to access and work with research data directly in Hubble.
Depending on the tools supported by your MCP client and your Hubble permissions, you can use Hubble MCP across much of the research workflow—from setting up studies and recruiting participants to analyzing results and creating reports.
Find and explore your studies
You can use your AI assistant to navigate your Hubble workspace and understand existing research.
For example, you can ask:
- “Show me my recent studies.”
- “Open the Discovery Interviews study.”
- “What is the research plan for this study?”
- “Who participated in this study?”
- “What tasks and questions are included?”
Your assistant can access information such as study metadata, research plans, tasks, participants, recordings, labels, reports, and collaborators.
Analyze research results
Hubble MCP can help analyze both quantitative and qualitative research data.
You can ask questions such as:
- “Summarize the results of this study.”
- “What are the biggest differences between these two surveys?”
- “Analyze the first five interviews and identify emerging themes.”
- “Which questions had the strongest responses?”
- “Show me examples from participants who mentioned pricing.”
- “Review this participant and check whether their responses appear low quality.”
For quantitative studies, your assistant can work with aggregated results such as distributions, completion rates, rankings, card sort results, tree testing paths, and usability outcomes.
For qualitative research, it can analyze open-ended responses, interviews, recordings, and participant-level responses.
Work with notes, labels, and themes
You can also use Hubble MCP to support qualitative coding and synthesis.
Your assistant can:
- View existing labels and themes
- Create new labels
- Add notes to specific moments in interview recordings
- Apply labels to notes
- Create themes
- Add supporting evidence to themes
- Review all evidence associated with a theme
For example:
“Create notes for moments where participants discuss trust, transparency, or pricing.”
Or:
“Create a theme around difficulty comparing lenders and add supporting interview evidence.”
Recording notes can be anchored to specific moments in a transcript, helping you maintain traceability between findings and source evidence.
Create and edit studies
Hubble MCP can also perform study authoring actions.
Your assistant can help you:
- Create a new study
- Write or update the research plan
- Add or modify study tasks
- Change certain study settings
- Publish a study
- Pause, resume, or archive studies
Supported research activities can include:
- Multiple-choice and open-ended questions
- Rating scales and matrix questions
- Video or voice responses
- Card sorting
- Tree testing
- Ranking exercises
- Five-second tests
- Live-site usability tests
- Prototype tests
- AI-moderated interviews
For example:
“Create a usability study to evaluate this onboarding prototype.”
Before creating a study, your assistant may ask about your research goals, target participants, key questions, and how the findings will be used so it can build an appropriate study design.
Create AI-moderated interviews
Hubble MCP can help configure AI-moderated interview studies.
You can define:
- Interview instructions
- Primary questions
- Follow-up questions
- Probing depth
- Voice or text-based interviews
- Interview language
- Whether the AI interviewer can reference earlier participant responses
For example:
“Create an AI interview exploring how UX researchers evaluate agentic products. Probe deeply into their workflow, tools, and challenges.”
Manage moderated research
For moderated studies, your AI assistant can also help manage scheduling and study settings.
Depending on your configuration, you can:
- View connected calendars
- Review scheduled sessions
- Change session duration
- Configure recurring availability
- Set specific interview times
- Update meeting or conferencing settings
For example:
- “Make these interviews 45 minutes.”
- “Show me the interviews scheduled this week.”
- “Set availability Tuesday through Thursday from 1–5 PM.”
Some scheduling changes require your explicit confirmation before they are applied.
Set up recruitment and review participants
Hubble MCP can support participant recruitment workflows as well.
Your assistant can help configure recruitment criteria such as:
- Country
- Age
- Gender
- Industry
- Job title
- Company size
- Professional skills
- Consumer interests or topics
It can also review incoming applicants and their screener responses.
For example:
“Review the applicants for this study and recommend who I should recruit.”
Your assistant can evaluate candidates and recommend whether to invite, reject, or skip them. Actions that affect real applicants require your approval before they are executed.
Create and save research reports
After analyzing your research, your assistant can create a report and save it directly to the study's Reports tab in Hubble.
For example:
“Analyze these interviews and create an emerging findings report.”
You can review and refine the findings with your AI assistant, then ask:
“Save this report to Hubble.”
Reports can include findings, supporting evidence, summaries, and tags so that you and your team—or another AI agent—can find the analysis later.
Example end-to-end workflow
With Hubble MCP, an AI assistant can support a workflow such as:
Open a study → review the research plan → analyze responses → identify patterns → pull supporting evidence → create notes and labels → develop themes → write findings → save the report to Hubble.
It can also support study operations:
Define the research goal → create a study → build the research plan → add tasks → configure recruitment → review applicants → launch the study → analyze results.
What actions require confirmation?
Hubble MCP includes safeguards for actions that can affect participants, studies, or research data.
Your AI assistant may ask for explicit approval before actions such as:
- Publishing a study
- Resuming participant access
- Inviting or rejecting recruitment applicants
- Replacing interview availability
- Switching study calendars
- Deleting a study
Read-only tasks such as viewing studies, reviewing results, and analyzing responses generally do not require confirmation.
Availability of MCP Actions
The exact actions available depend on your Hubble role (e.g., Admin has more control than a viewer), the MCP client you are using, and the tools currently exposed through Hubble MCP.
You can ask your AI/LLM:
“What Hubble MCP tools do you currently have access to?”
to see what actions are available in your current session.