The Retently MCP connects your Retently data to your AI agent, so you can read and act on survey responses, customers, campaigns, and reports in plain language, without logging in to Retently. It works with Claude Cowork, Claude Code, and OpenAI Codex.
This article is the starting point. It points you to the detailed guides for connecting and using the MCP, in the order most people need them.
Step 1: Connect the MCP to Your AI Agent
The default way to connect is through OAuth: there is no token to generate and nothing to paste. Pick the guide for the agent you use:
If OAuth is not an option for your client, or you need a connection with scoped, limited access (for example, read-only or no access to personal data), use a personal access token instead:
Before You Connect: Set Your Data Access
Open your AI Connections page in Retently. Under Customer Personal Identifiable Information Access, two toggles control whether your AI agent can read personal data and manage it. They are off by default.
With them off, the agent works with anonymous, quantitative data only. It cannot use filters, read contact properties, attribute responses to specific customers, or perform write actions. Unless you need to keep your data anonymous, enable both before you connect. If you already connected with them off, switch them on and reconnect for the change to take effect.
Each connect guide above covers this in detail.
Step 2: See What You Can Do
Once connected, the best first step is to ask your agent to list its tools:
What can the Retently MCP do? List all the tools and give a short description of each.
The MCP includes more than 50 tools across feedback, customers, campaigns, reports, and account usage. You do not need to memorize them or name a specific tool. Just ask in plain language, for example:
Pull all survey responses we received today.
How many responses did we receive this month so far, and how many are from detractors who mentioned pricing?
What trends do I have in Retently? Build a report for the NPS by Department trend.
For the full list of tool categories and more examples, see What the Retently MCP Can Do. For a complete reference of every tool with descriptions, plus a glossary of Retently terms, see Retently MCP Tools and Glossary.
Tell Us What Is Missing
The MCP is still being actively improved. If you run into a use case that is missing or could work better, let us know, ideally with the exact prompt you sent to your AI agent, so we can optimize for it.
Related