stigmi MCP Server — Connect Your Forms to AI Assistants Like Claude

What is MCP (Model Context Protocol)?

The Model Context Protocol, or MCP, is an open standard that lets AI assistants connect to external tools and services. Think of it as a universal adapter: instead of building custom integrations for every AI platform, a service can expose its capabilities through MCP, and any compatible AI assistant can use them immediately.

When an AI assistant like Claude has access to an MCP server, it gains the ability to call specific tools — creating records, reading data, updating settings — as part of a natural conversation. You describe what you want in plain English, and the AI translates that into the right tool calls behind the scenes.

The stigmi MCP server exposes 58 tools across 16 categories, covering everything from form creation and question management to response analytics, conditional logic, A/B testing, and CSV exports. All of these are accessible through a single stdio connection that any MCP-compatible client can use.

Why connect your forms to AI?

Managing forms through a dashboard works fine when you have a handful of them. But as your usage grows — more forms, more responses, more questions to iterate on — the overhead of clicking through a UI adds up. An AI-powered workflow eliminates the repetitive parts.

Automate form creation. Instead of manually adding questions one by one, describe the form you need: "Create a customer feedback survey with name, email, a 5-star rating, and an open-ended comments field." The AI creates the form, adds all the questions, and activates it — in seconds.

Query responses conversationally. Ask "How many responses did the onboarding form get this week?" or "What is the average NPS score?" and get instant answers without opening a spreadsheet or writing SQL.

Bulk operations. Need to update the branding on ten forms? Change the thank-you message across all your active surveys? Describe the change once, and the AI handles the rest.

Export and analyze. Pull CSV exports of response data, check drop-off analytics, or review A/B test results — all from within your conversation with Claude.

Getting started with the stigmi MCP server

Setting up the MCP server takes just a few minutes. You need Node.js 20 or later installed on your machine.

1. Build the server — Clone the stigmi repository, then run cd mcp && npm install && npm run build. This compiles the TypeScript source into a ready-to-run JavaScript file.
2. Configure Claude Desktop — Open your claude_desktop_config.json and add the stigmi MCP server entry with your credentials and the path to the compiled server file.
3. Set your credentials — Provide your stigmi email, password, and optionally an API token or admin secret as environment variables in the configuration.
4. Restart Claude — After saving the configuration, restart Claude Desktop. The stigmi tools will appear automatically in your tool list.

Here is what the configuration looks like:

{
  "mcpServers": {
    "stigmi": {
      "command": "node",
      "args": ["/path/to/stigmi/mcp/dist/server.js"],
      "env": {
        "STIGMI_BASE_URL": "https://stigmi.app",
        "STIGMI_EMAIL": "your@email.com",
        "STIGMI_PASSWORD": "your-password"
      }
    }
  }
}

If you use Claude Code instead of Claude Desktop, the project includes a pre-configured mcp.json at the repository root. Just edit it with your credentials and start Claude Code from the project directory — it auto-discovers the configuration.

What you can do with the MCP server

The stigmi MCP server covers the full form lifecycle. Here are the key capabilities organized by category:

Forms. Create, update, activate, pause, delete, and duplicate forms. Update settings like titles, descriptions, brand colors, welcome screens, scheduling, response limits, and custom CSS.

Questions. Save and replace all questions on a form in a single call. All 23 question types are supported, including short text, long text, multiple choice, dropdown, rating, NPS, file upload, payment, and more.

Responses. List all responses with their answers, export everything as CSV, and add annotations (notes) to individual responses for team collaboration.

Analytics. View totals, response timelines, drop-off rates per question, NPS scores, and A/B test results — all accessible through a single tool call.

Conditional logic. Create jump and end rules with conditions like "if rating is less than 3, jump to the follow-up question." You can even test your logic with sample answers before activating the form.

Custom endings and A/B testing. Create multiple thank-you screens with redirects and scoring, and set up A/B test variants with traffic weighting to optimize your forms.

Example workflows

To see how this works in practice, here are three real workflows you can run entirely through conversation with Claude.

1. Create a feedback form in under a minute.

Tell Claude: "Create a customer feedback form with fields for name, email, a 5-star service rating, a multiple-choice question about how they found us, and an open comments field. Then activate it." Claude calls create_form, then save_questions with all five questions, and finally update_form to set the status to ACTIVE. You get back a shareable link like stigmi.app/f/xK9mP2 — ready to send to customers.

2. Check NPS scores and drop-off rates.

Ask Claude: "Show me the analytics for my onboarding survey — I want NPS scores and where people are dropping off." Claude calls list_forms to find the right form, then get_analytics to pull the data. It returns a summary like: "Your NPS is 42. The biggest drop-off is at question 4 (address field) where 23% of respondents leave."

3. Export responses for a date range.

Tell Claude: "Export all responses from my product survey as CSV." Claude calls export_responses_csv and returns the data directly in the conversation. You can ask follow-up questions like "How many people rated us 5 stars?" and Claude will parse the data to answer. For API token users, the api_list_responses tool also supports date filtering with since and until parameters.

Authentication: session, API tokens, and OAuth

The MCP server supports three authentication modes, so you can choose the one that fits your setup.

Session authentication is the simplest option. Provide your stigmi email and password in the MCP configuration, and the server handles login automatically. It fetches a CSRF token, authenticates against the NextAuth endpoint, and stores the session cookies in memory. If the session expires, it re-authenticates transparently. This works with any plan, including FREE.

API tokens are available on the Pro+ plan. Generate a token from the Tokens page in your stigmi dashboard, then set it as the STIGMI_API_TOKEN environment variable. API tokens use Bearer authentication and give you access to the full API v1 tool set — listing forms, creating forms, managing questions, reading responses, and exporting CSV. Tokens have the format stg_... and can be revoked at any time.

OAuth2 is the recommended approach for production integrations. Create an OAuth client, obtain access and refresh tokens, and the MCP server handles token renewal automatically. This provides the most secure and flexible authentication flow, working with any plan level.

Get started with AI-powered forms

The stigmi MCP server is open source and ready to use today. Create your free stigmi account, build the MCP server with a single command, and connect it to Claude Desktop or Claude Code. Within minutes, you will be creating forms, querying responses, and exporting data through natural conversation.

Whether you are a developer building automated workflows, a team lead who wants to spin up surveys without leaving your AI assistant, or a power user looking to analyze form data faster — the MCP integration makes stigmi the most AI-friendly form builder available. All 58 tools, 23 question types, and the full analytics suite are at your fingertips.