SurveyNinja MCP server

The SurveyNinja MCP server and ready-made skills for your AI assistant. Describe a survey in plain words and the assistant builds it in your account, checks it before launch, reads the responses and prepares a report you can send.

Works in Claude, ChatGPT, Claude Code, Cursor, VS Code, Codex, Gemini CLI and other assistants that connect to remote MCP servers. Every SurveyNinja plan includes it, Free too.

You:  Build a survey for our repair shop customers: rate the work and collect
      contacts for a follow-up visit.
AI:   ✓ picked question types  ✓ built 7 questions and 2 branches  ✓ applied a theme
      The draft is ready and does not open by its link yet. Publish it?

Quick start

1. Have a SurveyNinja account. Sign up for free if you don't have one.

2. Add the server https://mcp.surveyninja.io/mcp/v1 to your assistant. Steps for each assistant are below.

3. Allow access. Sign in to SurveyNinja when the assistant asks, choose a workspace and click Allow access. Most assistants need no key; for the rest, use an API key as shown below.

4. Ask: “Show my surveys and how many responses came in this week.”

Claude Code: everything in one install

The plugin adds the MCP server and every skill at once:

/plugin marketplace add surveyninjaio/surveyninja-mcp
/plugin install surveyninja@surveyninja

Then type /mcp, select the SurveyNinja server and sign in. Only the server, without the skills:

claude mcp add --transport http surveyninja https://mcp.surveyninja.io/mcp/v1

Claude and ChatGPT

Claude on the web or in the desktop app: Settings → Connectors → Add custom connector. Enter a name, for example SurveyNinja, and the server address, click Add, then Connect, and allow access in SurveyNinja. On Claude Team and Enterprise an organization owner adds the connector first, then each member connects it.

ChatGPT: turn on developer mode in the ChatGPT settings (it is available on some ChatGPT plans), create a connector with the server address and OAuth authentication, then allow access in SurveyNinja.

Other assistants

One click:

Add the server to ~/.cursor/mcp.json for all projects, or to .cursor/mcp.json for one project, then open the MCP section in the Cursor settings and sign in to the surveyninja server:

{
  "mcpServers": {
    "surveyninja": {
      "url": "https://mcp.surveyninja.io/mcp/v1"
    }
  }
}

Add the server to .vscode/mcp.json in your project, start it from the file and sign in when VS Code asks. The server works in GitHub Copilot Chat in agent mode.

{
  "servers": {
    "surveyninja": {
      "type": "http",
      "url": "https://mcp.surveyninja.io/mcp/v1"
    }
  }
}

As an extension, with the server and the skills at once:

gemini extensions install https://github.com/surveyninjaio/surveyninja-mcp

Or only the server, in ~/.gemini/settings.json:

{
  "mcpServers": {
    "surveyninja": {
      "httpUrl": "https://mcp.surveyninja.io/mcp/v1"
    }
  }
}

If Gemini CLI does not offer to sign in by itself, run /mcp auth surveyninja.

With an API key. Create one in SurveyNinja under Settings → API / MCP, then add the server to ~/.codex/config.toml:

[mcp_servers.surveyninja]
url = "https://mcp.surveyninja.io/mcp/v1"
bearer_token_env_var = "SURVEYNINJA_API_KEY"

Codex reads the key from the environment: export SURVEYNINJA_API_KEY=YOUR_API_KEY.

Any client that connects to remote MCP servers over Streamable HTTP can use the same address. If it supports MCP sign-in (OAuth), it opens SurveyNinja for you to allow access. If it does not, pass an API key in a header:

Authorization: Bearer YOUR_API_KEY

The same configs as files are in clients/, and every step with troubleshooting is in docs/connect.md.

Related MCP server: Qualtrics MCP Server

At a glance

What it can do

Surveys. Create a survey from scratch, from a template or with AI. Edit questions, logic, design and settings, then publish.

Responses. Read responses, tag them and keep the account clean.

Reports. Summary, report, report with filters and AI report. Exports to CSV, Excel, Word, SPSS and PDF, public links to share results.

Delivery. Notifications, webhooks, integrations and invitations to the team.

The full list by section is in docs/tools.md, starter phrases in docs/prompts.md.

Skills

A skill is a ready-made playbook for the assistant. You ask “sort out the open-ended answers”, and the skill tells it which tools to call, in what order and how to present the result. Skills are text instructions with no code: nothing gets installed or run. Each one is also on surveyninja.io/skills with examples of the result.

Build and share

Check before launch

Understand results

Deliver responses

Keep the account tidy

All skills in one archive: all.zip.

Installing skills without the plugin

One command for any assistant: skills finds the agents you have installed and asks where to put the skills.

npx skills add surveyninjaio/surveyninja-mcp

Or unzip an archive into the folder your assistant reads skills from and restart it: .claude/skills/ for Claude Code, .agents/skills/ for Codex and Cursor, .gemini/skills/ for Gemini CLI. More in skills/README.md.

Security

• With Allow access the assistant works in the one workspace you chose and loses access when you leave it or remove the connector. An API key stops working the moment you delete it in Settings → API / MCP.

• Tools that cannot be undone, such as deleting surveys or responses, carry the destructiveHint annotation, so your client can ask before it calls them. Read what the assistant is about to do before you confirm.

• Every skill has a “What not to do” section: things the assistant must not do without your word.

• There are no secrets in this repository. Configs use YOUR_API_KEY as a placeholder. Found something that looks like a real key, or a security issue? See SECURITY.md.

Links

• SurveyNinja

• MCP server page

• Skills catalog

• MCP and API documentation

• LinkedIn

License

MIT, see LICENSE. The license covers what is in this repository: documentation, configs and skills. The SurveyNinja service itself is provided under its user agreement.