Back
product mcp agents
APR 16

Talk to Your Data: The Postbox MCP Server

Postbox's MCP server lets AI clients query forms, inspect submissions, summarize patterns, analyze spam, translate messages, and draft replies.

👤
Postbox Team
· · 2 min read

The Postbox MCP Server allows AI clients to interact with your form data through the Model Context Protocol. This post provides the exhaustive technical specifications for the server’s tools and universal setup instructions.

Conversational Data Intelligence

Why use the Postbox MCP Server?

While a dashboard is useful for manual inspection, it is an inefficient interface for Data Reasoning. The MCP Server exposes your submissions as a Programmable Data Source, allowing your AI agent to fetch, reason about, and act on your data using structured tools.

Universal Connectivity

MCP is a protocol designed for cross-client compatibility. Unlike local MCP servers that require binary installations, Postbox operates as a Remote MCP Server. This means you can connect it to any MCP-enabled tool (Cursor, Windsurf, VS Code, Claude Desktop, etc.) simply by pointing it to our discovery endpoint.

1. The Discovery URL

Point your MCP client to the following URL to discover the Postbox toolset:

https://usepostbox.com/mcp

2. Authorization

Postbox uses a browser-based authorization flow. When you first connect your client, it will prompt you to authorize the connection. This ensures your form data remains secure and accessible only to your specific AI session.


Client-Specific Setup

Claude Desktop

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "postbox": {
      "url": "https://usepostbox.com/mcp"
    }
  }
}

Cursor / VS Code / Windsurf

Go to your MCP Settings and add a new “Remote Server”:

  • Name: Postbox
  • URL: https://usepostbox.com/mcp
  • Auth: Follow the browser prompt to link your account.

Exhaustive Tool Reference

The MCP server exposes the following 17 tools, mapped 1:1 to our secure API.

1. Form Exploration

  • forms_list: List all forms and their primary slugs.
  • forms_get: Inspect a form’s name, intent, and field schema.
  • forms_create: Provision a new form backend.
  • forms_update: Update schema or enable AI features (spam, translation).
  • forms_delete: Remove a form.

Example Prompt: “What forms do I have? Inspect the ‘product-feedback’ form and tell me what fields it expects.”

2. Submission Analysis

  • submissions_list: Query submissions with search and filtering.
  • submissions_get: Fetch the full JSON payload and metadata.
  • submissions_delete: Delete a submission record.

Example Prompt: “List the latest 10 submissions for ‘bug-reports’. Search for any that mention ‘login failure’ and summarize them.”

3. Destination Management

  • destinations_list: See where data is being routed (Slack, Webhooks, etc.).
  • destinations_create: Add a new notification channel to a form.
  • destinations_delete: Unlink a destination.
  • destinations_regenerate_secret: Reset a webhook’s signing secret.

4. Knowledge Base Operations

  • knowledge_bases_list: Enumerate your knowledge bases.
  • knowledge_bases_get: Read the full content of a KB.
  • knowledge_bases_create: Provision a new KB for smart replies.
  • knowledge_bases_update: Evolve your documentation content.
  • knowledge_bases_delete: Delete a knowledge base.

Talk to your data. Connect the Postbox MCP Server today and stop hunting for signals in a spreadsheet. 5,000 lifetime submissions and full MCP access included. No credit card required.

Have thoughts?
Or connect for more dispatches.