Connect Marvin to tools such as Claude, ChatGPT, Figma Make, Cursor, and more with MCP. Bring your customer insights into every AI tool, and send new findings back to Marvin.
Ask questions and access your customer knowledge where you work. Once you’ve analyzed the data and compiled an answer or report, you can share it with your team in Marvin.
What kinds of data can I add to Marvin from AI tools?
With Marvin MCP’s write tools, you can do the following:
Create a new project: Add files you can use to set up new studies
Add or edit a project overview: Update the project overview to add more context, which can also be used to create discussion guides.
Create a discussion guide: Draft a discussion guide and review it before you add it to a project.
Publish an Insight: Create a report with tables, callouts, images, embeds and citations, and publish it to Marvin.
How do I add data to Marvin from AI tools?
Before you can share data, you must grant your tool write access. If you’ve already connected you must disconnect Marvin from your MCP connectors and then reconnect.
For example, when you connect to Claude, you must check the box against "Perform actions on your behalf within your permission level" to grant it write access.
Once you’ve connected Marvin, you can prompt the write actions directly from your tool. You have control over the changes that are made. The AI cannot add or update data unless you explicitly mention it in your instructions.
For example, if you’re scoping a new feature for your product, you may pull data from public websites, customer requests and feedback in Marvin, analytics, segmentation etc. Once you’re ready to share it to Marvin for your entire team to access, you can try this prompt in your tool:
“Create a project for this and publish my write-up as an Insight in Marvin.”
Formatting for headings, tables, checklists, callouts, images, links, and numbered citations is retained when
Next, once you’ve created gathered context about the project, you could add
“Update the project overview to explain the objectives, method, and metrics to all stakeholders.”
You could also prompt the AI tool to:
“Draft a discussion guide for interviews with users who have used feature X in the last 6 months.”
How do I safeguard existing data in Marvin?
All permissions from Marvin transfer over to your tools. You can only add data or make changes to projects to which you already have access.
All new projects and Insights you create via MCP are private by default. They are not shared with other Marvin users until you explicitly add an instruction to publish an Insight or share the project with Admins/editors/all users. You can also control these settings from Marvin. You have complete control over how and when data is shared with users.
An AI tool cannot delete data by itself. You must clearly mention you want to edit or update a project overview, discussion guide or Insight before the AI makes the change.
FAQs
Q. What data can I add to Marvin from AI tools?
A. You can add and update project overviews, create projects, draft new discussion guides, and create Insight reports.
Q. Can I create a study or record a meeting through Marvin MCP?
A. No, at the moment, you cannot create a study in Marvin from an AI tool.
Q. Can I disable write permissions for one or all tools I connect to Marvin?
A. Yes, AI tools cannot make changes to data in Marvin unless you grant write access while adding Marvin as a connector. With read-only access, you can still surface relevant insights in the tools where you work.
Q. Can I create labels in Marvin from my AI tool?
A. No, at the moment, you cannot add or update labels in Marvin via MCP.
Q. Can Admins restrict write access for team members?
A. No, currently, every user must grant write access for their own account.